EFICACIA DE MEDIDAS

COMPENSATORIAS PARA LA

CONSERVACIÓN DE ESTEPAS

AGRÍCOLAS EN ÁREAS IMPORTANTES

PARA LAS AVES DEL CENTRO

PENINSULAR

CARLOS PONCE CABAS

TESIS DOCTORAL

2016

A mi familia y amigos

A Arantza

UNIVERSIDAD AUTÓNOMA DE MADRID

FACULTAD DE CIENCIAS

DEPARTAMENTO DE ECOLOGÍA

EFICACIA DE MEDIDAS COMPENSATORIAS PARA LA CONSERVACIÓN DE ESTEPAS AGRÍCOLAS EN ÁREAS IMPORTANTES PARA LAS AVES DEL CENTRO PENINSULAR

Memoria presentada por Carlos Ponce Cabas para optar al grado de Doctor en Ciencias Biológicas

Bajo la dirección de:

Juan Carlos Alonso López

Profesor de Investigación

Dpto. Ecología Evolutiva

Museo Nacional de Ciencias Naturales

Luis Miguel Bautista Sopelana

Científico Titular

Dpto. Ecología Evolutiva

Museo Nacional de Ciencias Naturales

Madrid, 2015

“Hemos reseñado el daño que causa la connotación negativa de los espacios

esteparios españoles de singular belleza e interés científico…, importantes

recursos naturales que merecen urgente protección y mejor conocimiento de

quienes, mirando sin ver, los menosprecian.”

Fernando González Bernáldez

Índice

Agradecimientos 1

Introducción general 7

Capítulo 1. Carcass removal by scavengers and search accuracy affect bird mortality estimates at power lines 23

Capítulo 2. Wire marking results in a small but significant reduction in avian mortality at power lines: A BACI designed study 53

Capítulo 3. Effects of organic farming on plant and arthropod communities: a case study in Mediterranean dryland cereal 83

Capítulo 4. Effects of agri-environmental schemes on farmland birds: Do food availability measurements improve patterns obtained from simple habitat models? 119

Capítulo 5. Effects of farming practices on nesting success of steppe birds in dry cereal farmland 163

Discusión general 193

Conclusiones 207

Anexo 1. Programa de Medidas Agroambientales del Área Importante para las Aves Talamanca-Camarma

209

1

AGRADECIMIENTOS

He intentado que en este apartado aparezcan todas las personas que han hecho

posible que esta tesis llegue a buen puerto, aunque es probable que me deje a

algunas. Espero que sean pocas y no me lo tengan en cuenta. Simplemente no

puedo acordarme de todas las personas que han participado, con su apoyo,

conocimientos o en cualquier otro aspecto.

Recreación de conversación mantenida con mi primo Jandro, el gallego,

hace un eón de años. Probablemente no fuera así exactamente, pero a mí me gusta

recordarla de esta manera.

Jandro: primo, vamos a ir a ver pájaros en una laguna de mi instituto. ¿Te

vienes?

Yo: ¿Ver pájaros? ¿Qué es eso?

Jandro: pues vamos con unos prismáticos y buscamos pájaros para saber de

qué especie son.

Yo: pues es bastante raro lo que dices, pero os acompaño que si no tu padre

me pondrá a cortar zarzas del camino.

Un rato después...

Jandro: ¡mirad, allí hay una polla de agua!

Yo: jajajaja. ¿Una polla? Jajajajaja.

Jandro: fíjate, tiene el pico rojo y parece una gallina. Míralo en la guía...

Desde aquel mismo momento en que miré la foto y vi el "pájaro" ese del que

hablaba mi primo supe a qué quería dedicar el resto de mi vida. Me pareció

maravilloso -y de locos- saber que "no todos los pájaros son iguales" y que alguien

se había dedicado a sacar fotos de estos animalillos para publicarlas en un libro.

Así que primo jandro, tienes que saber que mi primer agradecimiento es para ti.

Gracias a ese momento en la lagunilla de tu instituto soy ahora lo que soy, un

"bichólogo". Tú eres Doctor en biología y, aunque no somos de la misma rama,

sabes lo que cuesta desarrollar una tesis.

2

Han pasado muchos años desde entonces y muchas cosas han cambiado en

mi vida, pero acabar esta tesis es, probablemente, lo mejor que me ha pasado a

nivel profesional.

Es obligatorio agradecer el enorme trabajo realizado por mis directores de

tesis, Juan Carlos Alonso y Luis Miguel Bautista. Os agradezco a ambos la paciencia

que habéis tenido conmigo en algunos momentos. Desde luego, lo que sé sobre

ciencia es, básicamente, gracias a vosotros dos. Juan Carlos me permitió

incorporarme al Museo Nacional de Ciencias Naturales con una beca para hacer

trabajo de técnico. En ese momento no me creía que fuera a trabajar en tal

institución, y encima con avutardas. Solicitamos varias becas externas al equipo

para hacer la tesis, pero me rechazaban por no tener un expediente "adecuado".

Esta tesis es la muestra de que se equivocaron. Aunque la beca, y posterior

contrato, que me proporcionó Juan Carlos, estaba ligada a trabajo de técnico, me

permitió compaginarlo con la tesis, no sin grandes esfuerzos. Ambos han hecho

posible el comienzo y la finalización de esta tesis. A pesar de todo el trabajo de mis

directores quiero aclarar que cualquier error que se encuentre en esta memoria de

tesis doctoral es únicamente achacable a mí.

Mario Díaz me asesoró a la hora de decidir qué tipo de muestreo era más

conveniente hacer en el campo para las semillas y artrópodos, una de las bases

fundamentales de esta tesis. Me informó de un curso sobre medidas

agroambientales que se iba a hacer. Ese curso me fue de gran utilidad para

desarrollar diferentes aspectos de la tesis. Espero que haya sabido plasmarlos de

una forma adecuada.

Quiero agradecer la labor de Javier Seoane, por acceder a ser mi tutor en la

UAM y por su ayuda con algunos de los análisis llevados a cabo en esta tesis. Nos

conocemos desde hace muchos años y es un referente para mí en el tema de las

aves y la investigación.

Desde luego, esta tesis nunca habría podido llegar a buen término sin la

participación de los agricultores en el Programa de Medidas Agroambientales. A

pesar de algunos momentos complicados, siempre habéis hecho lo posible para

que nuestros experimentos pudieran llevarse a cabo.

3

Quiero agradecer también a los primeros avutarderos con los que trabajé en

el museo. A Carlos Martín, Carlos Palacín, Marina Magaña, Beatriz Martín y Pablo

Sastre por los ratos que pasamos juntos durante años, con las avutardas y con

otras especies. Algunos de vosotros habéis colaborado en gran medida en varios

capítulos de la tesis. Carlos Martín y Carlos Palacín hicieron posible la señalización

de los tendidos eléctricos y Beatriz los recorrió en búsqueda de cadáveres. Marina

fue la responsable del Programa de Medidas Agroambientales hasta que yo asumí

esa labor. Con Pablo compartí bastantes jornadas de campo registrando molestias

a las avutardas.

También me gustaría expresar mi gratitud hacia mis últimos compañeros en

el museo. Iván salgado y Aurora Torres, con los que he vivido inolvidables

momentos en el campo. Iván ha sido mi compañero de andanzas en el capítulo "de

los huevos" y Aurora siempre me ha ayudado con el R y el GIS, desde que un

revisor me pidió que hiciera un modelo mixto -me sonaba a un sandwich- en lugar

de una prueba de datos pareados, hasta enviarme comandos de R. Aurora y yo

hemos compartido muchas horas de censo y de capturas, mucho más agradables

que si fuera yo solo. Estoy seguro de os espera a ambos un futuro muy bueno en el

mundo científico.

Varios estudiantes han participado de forma notable durante diferentes

fases de esta tesis. Quiero agradecer especialmente su colaboración a Elena,

Gonzalo, Natalia, Desi, Dácil, Iris y Alberto.

Un agradecimiento especial debe ir a Carolina Bravo, con quien he

compartido muchas horas caminando por el campo con una cinta métrica atada a

la espalda o hablando de "las lolas". Recuerdo sobre cualquier otra cosa las

quedadas con el todoterreno y los calores que pasamos en algunos lugares, como

en Aranjuez, que desembocaron en problemillas leves de salud. Gracias a tu

aportación se han podido llevar a cabo casi todos los capítulos de esta tesis (y

algunos artículos que vendrán después). Espero poder seguir colaborando contigo

en el futuro.

No puedo olvidarme de Rafael Barrientos y de Jose Manuel Álvarez. Cada

uno habéis estado ahí en diferentes momentos. Con Rafa he compartido trabajo de

4

campo y gabinete durante años, mucho más tiempo de lo que duró su corta

vinculación al equipo. Jose me ayudó en algunos momentos de la tesis, con análisis

variados. He aprendido mucho de ti. Desde luego, con ambos he pasado también

momentos "distendidos" dentro y fuera del museo.

Antes de entrar a trabajar en el Museo Nacional de Ciencias Naturales pude

formarme como biólogo y como ornitólogo. Debo agradecer a un profesor en

particular la forma de enseñar y su accesibilidad hacia los alumnos. Como no podía

ser de otra forma, me refiero a Quico. La asignatura que él impartía era una

maravilla. Cuando le preguntábamos, él parecía encantado, así que nosotros

seguíamos preguntando, con ciertas repercusiones sobre el temario, pero mereció

la pena. Las salidas de los cursos de doctorado eran "especiales", con grandes

variaciones sobre lo previsto. Gracias Quico por ofrecerme en su momento un

contrato para trabajar con las alondras de dupont (ricotí no me gusta). Tú me

animaste a pedir la beca que había salido en el museo para trabajar con las

avutardas. Así lo hice, y te lo agradeceré siempre. Recuerdo también cuando te dije

mi apellido. En ese momento te cambió la cara. Eso de saber que tú y mi padre

trabajasteis juntos en tus comienzos me hizo mucha ilusión. Quico, tu legado en

forma de publicaciones y actuaciones diversas para conservar este grupo de

"pájaros marrones" al que llamamos esteparias es encomiable. Ningún amante de

las aves podrá olvidarte.

Mi formación como pajarero empezó (y continua actualmente) de la mano

de Monticola. Con ellos he aprendido lo que sé sobre aves. Quiero agradecer

especialmente a Juancho sus horas "no muertas" hablando de aves, y anillando

para sacar fichas de muda que se tradujeron en diversos trabajos. ¿Cuántas aves

hemos anillado tú y yo juntos? Es una lástima que ahora estés tan lejos. También es

necesario dar las gracias a otros monticoleños (o monticolocos) con los que

también he disfrutado mucho. Cristian, Óscar, Virginia, con vosotros he compartido

innumerables horas de campo, de día y de noche. Me alegro de que todos forméis

parte de mi vida.

Mis años en el museo han sido muy especiales por toda la gente que he

conocido. Cada uno dedicado a un tema diferente, pero entre todos hacemos un

grupo "curioso". Quiero agradeceros a todos vuestra amistad, que espero dure

5

hasta que las ranas tengan melena. Seguro que me olvido de algunos, pero muchas

gracias a Rigo, Jose, Chechu, Diego, Elisa, Chio, Dani, Marcos, Marti, Sergio, Jimena,

Reimon, Jaime, Eva, Roger, Shirin, Geizi, Natalia y el resto de pestuz@s del museo.

¡¡Que siga la fiesta!!!

Gracias a mis compañeros de la 1111, con los que he tenido muchos

momentos inolvidables dentro del despacho, incluida alguna que otra fiestecilla.

Antón, Ibáñez, Isaac, Juan, Octavio, Rafa, Cantarero (como si fueras de la 1111) y,

por qué no decirlo, Regan, entre todos habéis hecho que mi estancia en ese

despacho haya sido muy agradable.

No quiero olvidarme de mi familia. Mis hermanos y mis padres, entre todos

me habéis animado a hacer lo que siempre quise, trabajar con aves.

Por último, quiero agradecer el apoyo a la persona más importante de mi

vida, Arantza. Me has apoyado y animado en momentos muy duros y hemos

compartido nuestra pasión por las aves. Me has acompañado al campo, a echar

horas muestreando, censando y con los documentos finales de esta tesis. Siempre

has estado ahí, conmigo. ¡¡GRACIAS!!

6

7

INTRODUCCIÓN GENERAL

Las estepas son áreas abiertas con vegetación escasa y generalmente de porte bajo

(Valverde 1958). Sin embargo, con frecuencia también se denominan

pseudoestepas o estepas agrícolas o cerealistas al complejo de hábitats agrarios de

cereal de secano de gran extensión. La variación en el nombre se produce para

diferenciarlas de las verdaderas estepas del este de Europa, que no tienen

aprovechamiento agrícola. En todos los casos estos espacios poseen las siguientes

características:

Simplicidad estructural: Tienen una cubierta vegetal que incluye solamente

herbáceas o arbustos de pequeño porte.

Buena visibilidad, como consecuencia de la escasez de vegetación de gran

porte

Ausencia de lugares de nidificación protegidos (como árboles o acantilados)

Necesaria exposición a las inclemencias climáticas, debido a la escasez de

refugio: viento, lluvia, insolación, etc.

Altas fluctuaciones en las temperaturas entre diferentes estaciones, y

especialmente entre el día y la noche

Escasez de zonas provistas de agua de manera permanente

Gran dinamismo, puesto que se producen grandes modificaciones del

paisaje en poco tiempo

En la Península Ibérica podemos distinguir tres grandes unidades (Suárez et

al. 1991): las estepas leñosas, dominadas por arbustos de pequeño porte,

fundamentalmente caméfitos, entre los que suceden frecuentes calveros de suelo

desnudo. En esta categoría también se incluyen las comunidades halófilas con

elementos subarbustivos, espartales y albardinales. Por otro lado, se encuentran

los pastizales, llanuras densamente cubiertas de comunidades herbáceas entre las

que se intercalan zonas de matorral más alto y pies aislados de arbolado.

Finalmente, la estepa cerealista, que se compone de territorios llanos o

ligeramente ondulados dedicados en su mayoría al cultivo de cereal en secano.

8

Con la denominación de aves esteparias se conoce a un grupo de aves que

tienen en común el desarrollo de la totalidad o parte de su ciclo biológico en un

hábitat determinado, el medio estepario. Estas aves son un grupo heterogéneo

desde un punto de vista filogenético y sistemático que, sin embargo, presentan

similares adaptaciones morfológicas, fisiológicas, etológicas y ecológicas que

favorecen su supervivencia en este medio y que resultan ejemplos de convergencia

adaptativa. La presencia de aves esteparias en nuestra región data de antiguo, y

está documentada en el registro fósil (Santos & Suárez 2005).

Estas especies de aves han estado capacitadas para su expansión,

independientemente de su origen, gracias a una mayor dependencia de la

estructura del paisaje frente a caracteres edáficos o florísticos, lo que justifica la

colonización de medios tan humanizados, especialmente los sujetos a regímenes

de explotación tradicionales y cierta diversidad paisajística. El cultivo de cereal de

secano ofrece mejores condiciones de termorregulación para la nidificación de

algunas aves esteparias que las ofrecidas por la vegetación natural de las estepas

(Farago 1986).

Para avalar la importancia que tienen estos paisajes agrícolas para la

avifauna ibérica cabe decir que el 27% de las IBA —acrónimo inglés de Áreas

Importantes para las Aves— de nuestro territorio presentan medios típicamente

agrícolas y que la Península Ibérica tiene la mejor representación de aves

esteparias de la Unión Europea, donde los agrosistemas ocupan casi la mitad de su

superficie (European Comission, 2011). En la Península Ibérica destaca la

importancia de cultivos cerealistas de secano, presentes en un 22% de las IBA, que

resultan esenciales para el desarrollo de gran parte del ciclo biológico del muchas

especies de aves. Dichos sistemas agrícolas están muy ligados al ser humano y por

ello están expuestos a ciclos muy dinámicos que implican alteraciones muy

acentuadas sobre el entorno en períodos de tiempo muy cortos como la recogida

del cereal, laboreo de las tierras, etc. En definitiva, no es extraño que la adecuada

conservación de los ecosistemas peninsulares pase por un apropiado

mantenimiento de los sistemas agrarios tradicionales, donde el ser humano no sólo

no es dañino, sino que es fundamental para su conservación.

9

Actualmente, los taxones ligados a estos medios se enfrentan a problemas

de diversa índole que se pueden agrupar en la destrucción, alteración y

fragmentación de su hábitat. Es necesario recordar que estas amenazas genéricas

son la principal causa de pérdida de biodiversidad a nivel mundial no sólo en las

estepas de otros continentes, sino en la mayoría de los ecosistemas del planeta. La

multitud de problemas de las estepas superan las posibilidades temporales y

materiales de una tesis doctoral para ser tratados adecuadamente. En la presente

tesis nos centraremos en dos aspectos relacionados con los impactos en las aves

esteparias: la mortalidad de aves en tendidos eléctricos y la intensificación

agrícola.

Problemática asociada a los tendidos eléctricos

El transporte de electricidad desde las plantas a los usuarios se produce

habitualmente vía aérea mediante líneas eléctricas. Sin embargo, éstas provocan la

muerte de gran cantidad de aves, tanto por electrocución como por colisión,

además de otros problemas de diversa índole (Ferrer 2012). En los Estados Unidos

de América se estima que mueren más de 175 millones de aves como consecuencia

de las colisiones y electrocuciones en tendidos eléctricos (Manville 2009).

En España, aunque se desconoce la magnitud real del problema a escala

nacional, sí se sabe que las muertes por colisión contra tendidos eléctricos

suponen un grave problema, para las aves en general y para las esteparias en

particular (Alonso et al. 1994, Palacín et al. 2004) y, en 2008 se elaboró un Real

Decreto (Real Decreto 1432/2008), por el que se establecen medidas para la

protección de la avifauna contra la colisión y la electrocución en líneas eléctricas

de alta tensión.

Además, las colisiones contra tendidos eléctricos son la principal causa

conocida de mortalidad no natural en la Avutarda común (Otis tarda, Palacín et al.

2004), la especie más representativa de las estepas. Dichas muertes tienen un

grave impacto en la dinámica poblacional de la especie, hasta el punto de que es el

principal factor influyente en la probabilidad de extinción de las avutardas en

varios lugares del centro peninsular (Martín 2008). Sin embargo, el impacto de los

tendidos eléctricos no se traduce exclusivamente en la muerte directa de las aves

10

colisionadas o electrocutadas, sino que también pueden cambiar patrones

comportamentales (Sergio et al. 2004).

Desde hace años, las investigaciones sobre la viabilidad poblacional de las

aves afectadas han tratado de mitigar las muertes de aves mediante medidas

correctoras. Para el caso de la colisión, la medida más habitual es la incorporación

de dispositivos anticolisión en los cables, ya sea el cable de tierra o en los propios

conductores (Ferrer 2012). El hecho de colocar los dispositivos en el cable de

tierra se debe a que se le considera responsable de la mayor parte de las colisiones

debido a su menor diámetro (Heijnis 1980, Beaulaurier 1981). Estos dispositivos

aumentan la visibilidad de los cables, de manera que las aves puedan evitar la

colisión.

Sin embargo, existen pocos estudios diseñados correctamente para obtener

conclusiones sobre la eficacia de los dispositivos. Los estudios suelen carecer de un

tamaño muestral suficiente o de un correcto diseño experimental (Barrientos et al.

2011). Además, la mortalidad real producida se subestima al no incorporar

factores externos a las propias colisiones que, sin embargo, influyen en la estima

de su frecuencia (Bevanger 1999). Dichos factores son:

1.- la detectabilidad debido a la experiencia de los observadores y

características del hábitat

2.- imposibilidad de muestrear en determinados lugares

3.- la desaparición de los cadáveres colisionados por la acción de animales

carroñeros

Problemática asociada a la intensificación agrícola

La agricultura ha experimentado tras la Segunda Guerra Mundial un notable

proceso de intensificación en la mayor parte de los países Europeos (Gardner

1996, Robinson & Sutherland 2002). Dicha intensificación se define como el

aumento de la producción agrícola por unidad de superficie (Donald et al. 2001).

Para lograr ese incremento en la producción, la agricultura ha cambiado

notablemente (Sans et al. 2013). Así, nos encontramos en la actualidad con:

incremento en los insumos utilizados (fertilizantes sintéticos, biocidas), pérdida de

11

heterogeneidad de cultivos (tanto en especies como en variedades empleadas),

puesta en cultivo de terrenos incultos, concentración parcelaria (con la

consiguiente eliminación de bordes entre parcelas), un aumento en los trabajos de

laboreo en las zonas agrícolas (que dificultan la presencia de plantas silvestres,

invertebrados y fauna vertebrada), el laboreo de parcelas recién cosechadas,

descenso en la anchura de los bordes entre parcelas, siembra de variedades de

ciclo corto, y tendencia a cosechar en épocas más tempranas.

Las consecuencias directas de la intensificación agrícola han sido ciertos

costes en términos ecológicos. Fundamentalmente, se ha producido una gran

reducción de la biodiversidad en el medio agrícola de diferentes grupos

taxonómicos (Clough et al. 2005, Fuller et al. 2005, Donald et al. 2006), incluyendo

aves, mamíferos, artrópodos y plantas silvestres (ver revisión en Benton et al.

2003).

En particular, se han detectado disminuciones alarmantes en muchas

especies de aves que, gracias al seguimiento que hacen los países de sus

poblaciones, son buenas indicadoras de las variaciones en calidad del hábitat.

Según los resultados obtenidos por el European Bird Census Council (EBCC

2010), las aves ligadas a ambientes agrarios están en claro retroceso a nivel

Europeo. De las 36 especies de aves consideradas para el último cálculo de su

tendencia (EBCC 2014), 26 muestran una tendencia negativa. De ellas, 19 tienen

una tendencia de "declive moderado" (menor del 5% anual), y 3 un declive fuerte

(mayor del 5% anual). Según el EBCC, existe un declive de casi el 50% para el

conjunto de especies ligadas a medios agrarios.

Las investigaciones llevadas a cabo en varios países concuerdan con la

información aportada por el EBCC. Así, por ejemplo, en Reino Unido se detectó que

en 20 años (1970-1990) el 86% de las especies de aves (de 28 estudiadas) habían

reducido sus rangos de distribución, o que el 83% de 18 eran menos abundantes

(Fuller et al. 1995). No solo eso, sino que un estudio comparativo reflejó que los

tamaños poblacionales de aves eran, de media, el 52% entre los años 1968 y 1995

en algunas especies en declive (Siriwardena et al. 1998).

12

La situación en España no es más alentadora. De hecho, según Escandell et

al. (2011), algunas de las especies están decayendo incluso a un ritmo mayor que

en el conjunto de Europa. Así, nos encontramos con descensos acumulados del

26% (hasta un 72% en el norte de España) en la Calandria común (Melanocorypha

calandra), del 20% en la Alondra común (Alauda arvensis) o del 72% en la Ganga

ortega (Pterocles orientalis) desde el año 1998, momento en el que comenzó el

Programa de Seguimiento de Aves Comunes Reproductoras (SACRE) de

SEO/BirdLife.

Los motivos relacionados con la intensificación agrícola por los que la

tendencia de las aves ligadas a estos medios es tan negativa son el descenso en la

cantidad y calidad de alimento y la pérdida de lugares seguros para nidificar, con

aumento en la depredación de nidos (Newton 2004).

En respuesta a estos hechos, a lo largo de las últimas décadas se han

generalizado en varios países europeos diversos sistemas de primas a los

agricultores, a cambio de que estos modifiquen sus prácticas agrícolas y sigan unas

directrices tendentes a mejorar la conservación de la biodiversidad en el medio

agrícola (Programas de Medidas Agroambientales). Así, la Política Agraria Común

(PAC) de la Unión Europea introdujo en su reforma de 1985 la posibilidad de que

los Estados miembros invirtiesen fondos en la conservación de zonas

ambientalmente sensibles y, en 1992, se aprobó la regulación CEE 2078/92,

demandando la aplicación de medidas agroambientales. Por último, en las últimas

reformas de la PAC no se condiciona la cuantía de las compensaciones a la

producción agrícola, sino a la superficie. Además, se condicionan las ayudas

recibidas a unas normas y prácticas básicas (condicionalidad).

Sin embargo, tras un largo periodo de funcionamiento de estos programas

de medidas agroambientales, no existe consenso sobre el grado de cumplimiento

de su objetivo, el cual no es otro que contribuir a conservar la biodiversidad y

revertir los problemas generados por la intensificación agrícola. La ausencia de

conclusiones claras sobre la efectividad de los programas de medidas

agroambientales deriva del hecho de que muchas de las prácticas habituales en

conservación de la naturaleza se basan más en asunciones y conclusiones extraídas

sin fundamento o heredadas de anteriores experiencias que en un análisis

13

científico de la realidad. En la revisión de Kleijn y Sutherland (2003) sólo

encontraron 62 estudios de tan solo seis países en los que se examinara la eficacia

de los planes de medidas agroambientales. La mayoría de esos estudios habían

sido realizados en sólo dos países (Reino Unido y Holanda), sólo uno en Portugal, y

ninguno en Francia ni en España, circunstancia que llamó la atención de los

autores, quienes advierten reiteradas veces en su trabajo sobre la urgente

necesidad de estudios de este tipo en países del entorno mediterráneo.

Por último, en muchos de los estudios no se detectó beneficio alguno para la

biodiversidad tras la aplicación de las medidas agroambientales (Kleijn et al.

2006). Es más, en algunos (ver revisión en Kleijn & Sutherland 2003) se

observaron efectos negativos sobre algunos de los grupos analizados. Así pues, hoy

por hoy es cuestionable la eficacia de muchos de los programas de ayudas

agroambientales vigentes, e incluso, la conveniencia de que dichos programas

sigan implementándose en el futuro. Teniendo en cuenta que a lo largo de la

pasada década se han invertido sólo en Europa 24 millones de euros en programas

de ayuda agroambientales, es urgente la realización de un mayor número de

estudios que evalúen de forma científicamente contrastada la eficacia de dichas

medidas, para poder optimizar los recursos económicos que actualmente se

destinan a hacer compatible la agricultura y la conservación de la biodiversidad en

el medio agrícola.

En esta tesis se evalúa la eficacia del siguiente plan de medidas compensatorias:

Proyecto de Medidas preventivas, correctoras y compensatorias de la

afección de la M-50 y de la Autopista de peaje R-2 a la población de avutardas

y otras aves esteparias de la IBA Talamanca-Camarma, y al LIC Cuenca de los

ríos Jarama y Henares

En abril de 2001, la empresa Autopista del Henares, S.A. (HENARSA),

constructora de la R-2 Madrid-Guadalajara y del tramo de la M-50 con el que

enlaza, encargó a un equipo del CSIC dirigido por el Profesor de Investigación J.C.

Alonso el Proyecto de medidas preventivas, correctoras y compensatorias de la

afección de la M-50 (tramo M-607/N-IV, subtramo N-I/N-II) y de la Autopista de

14

peaje R-2 a la población de avutardas y otras aves esteparias de la IBA Talamanca-

Camarma, y al LIC Cuenca de los ríos Jarama y Henares.

La autopista de peaje R-2 atraviesa la zona sur de la ZEPA 139 "Estepas

cerealistas de los ríos Jarama y Henares" en la Comunidad de Madrid y el sureste

de la ZEPA 167 "Estepas cerealistas de la campiña" en la provincia de Guadalajara,

mientras que la M-50 se encuentra próxima a la ZEPA 139 y enlaza con la R-2

(Figura 1).

Figura 1. Zonas de especial Protección para las Aves (ZEPA) 139 y 167 y trazado de las autopista R-

2 y M-50. Elaborado por Aurora Torres y modificado parcialmente en esta tesis.

En el marco de dicho Proyecto se aplicó un plan de medidas compensatorias

de conservación para paliar el efecto de la construcción de ambas carreteras y

compensar a las zonas agrícolas afectadas por la pérdida directa de hábitat. Dicho

plan de medidas compensatorias comprende diferentes actuaciones que se pueden

agrupar en:

1.- Plan de divulgación y educación ambiental en diferentes centros educativos

(no tratado en la presente tesis).

2.- Actuaciones para reducir las molestias a la fauna y la mortalidad. Se llevó a

cabo la señalización de tendidos eléctricos en la Comunidad Autónoma de

Madrid y en la provincia de Guadalajara. Esta actuación se desarrolla en los

15

capítulos 1 y 2 de la presente tesis. Para el estudio de las molestias, se

realizó un trabajo específico, no incluido en esta memoria de tesis doctoral.

3.- Actuaciones para mejorar la calidad del hábitat. Se implementó un Plan de

Medidas Agroambientales en la IBA Talamanca-Camarma en las provincias

de Madrid y Guadalajara. Esta actuación se desarrolla en los capítulos 3, 4 y

5 de la presente tesis. Las medidas agroambientales implementadas en el

Programa fueron:

a) Mejora y mantenimiento del barbecho tradicional (mantenimiento

de rastrojeras)

b) Barbecho semillado con leguminosas

c) Retirada de la producción de tierras

d) Rotación de cultivos trigo- girasol

e) Cultivo de cereal no tratado

Las prescripciones concretas de cada medida se muestran en el Anexo 1 de

la tesis.

4.- Plan de seguimiento y valoración de las medidas. Comprende toda la tesis

en su conjunto.

Desarrollo de las actuaciones

El procedimiento efectuado para la señalización de los tendidos, consistió en

contactar con las empresas propietarias de cada uno de los tendidos en los que se

habían detectado importantes mortalidades de aves. Después del análisis de

viabilidad (realizado por la empresa propietaria del tendido) se produjo la petición

de ofertas a empresas fabricantes de dispositivos anticolisión. Una vez dichos

dispositivos fueron colocados se procedió a evaluar su eficacia.

El procedimiento para la selección de parcelas a incluir en el Programa de

Medidas Agroambientales se refleja en la figura 2.

16

Figura 2. Procedimiento resumido del Programa de Medidas Agroambientales.

Se llevó a cabo un primer contacto con las diferentes asociaciones de

agricultores en la zona de actuación. Una vez que los agricultores enviaban su

solicitud de incorporación de parcelas al Programa, ésta era evaluada por el equipo

dirigido por el Profesor de Investigación del CSIC Juan Carlos Alonso. En ese

momento se descartaban algunas parcelas, otras se aprobaban y otras se

rechazaban de forma condicional. En este último caso se volvía a contactar con los

agricultores para valorar la posibilidad de cambiar el tipo de medida

agroambiental a aplicar o la localización de la parcela solicitada y se evaluaba la

nueva propuesta. Si ésta satisfacía las necesidades del Programa se admitían en el

mismo. En caso contrario se rechazaban. Al implementar cada medida, se llevaban

a cabo controles periódicos sobre las parcelas. Si los agricultores cumplían el

compromiso adquirido se realizaba el pago de las primas correspondientes.

Objetivos y estructura de la tesis

El objetivo principal de esta tesis es evaluar la eficacia del plan de medidas

compensatorias. Concretamente, nos centramos en las actuaciones 2, 3 y 4

mencionadas anteriormente.

Así pues, la tesis doctoral está estructurada en los siguientes capítulos:

17

Capítulo 1. Carcass removal by scavengers and search accuracy affect bird

mortality estimates at power lines.

En este capítulo se desarrolla un experimento para cuantificar la

desaparición de cadáveres en tendidos eléctricos debida a la acción de animales

carroñeros, así como las diferencias en la detectabilidad de dichos cadáveres por

diferentes observadores que varían su grado de experiencia en localizar aves

muertas bajo los tendidos eléctricos. Ambas cuestiones son fundamentales para

obtener estimas reales del impacto que tienen las líneas eléctricas sobre las aves.

Capítulo 2. Wire marking results in a small but significant reduction in avian

mortality at power lines: A BACI designed study.

Mediante un diseño BACI se evalúa la eficacia de la señalización con

espirales salvapájaros de tramos de líneas eléctricas en la Comunidad de Madrid y

la provincia de Guadalajara que atraviesan zonas agrícolas de importancia para las

aves esteparias. También se estudia la posibilidad de que diferentes tamaños de

espiral produzcan resultados diferentes, así como la señalización de diferentes

tipos de tendidos eléctricos (de transporte o de distribución).

Capítulo 3. Effects of organic farming on plant and arthropod communities: A case

study in Mediterranean dryland cereal.

Se realiza un análisis de la eficacia del cultivo ecológico de cereal de secano

para plantas y artrópodos. Se comparan valores de abundancia, riqueza y

diversidad para ambos grupos entre diferentes formas de manejo, convencional y

ecológica. También se compara la biomasa de invertebrados obtenida en ambos

tipos de manejo.

Capítulo 4. Effects of agri-environmental schemes on farmland birds: do food

availability measurements improve patterns obtained from simple habitat models?

En este capítulo se estudia la respuesta de las aves ante el manejo de

parcelas incluidas en el Programa de Medidas Agroambientales. Concretamente, se

examina la importancia de la estructura del hábitat, el paisaje y la cantidad de

alimento presente en la abundancia, riqueza, diversidad y un Índice que incluye

información de la lista de Especies Amenazadas de Preocupación en Europa (SPEC)

18

desarrollada por BirdLife International. Asimismo, se compara la utilidad y

capacidad predictiva de dos tipos de modelos, uno en el que en incluyó la cantidad

de alimento (modelos costosos) y otro en el que se emplearon variables de

superficie (modelos sencillos).

Capítulo 5. Do agri-environmental schemes effectively protect nests from

predation? An experimental study.

En este capítulo se desarrolla un experimento de depredación de nidos. Se

relaciona las tasa de depredación de nidos con variables a escala de paisaje,

parcela y estructura de la vegetación en torno al nido. Se emplean cámaras de

fototrampeo para conocer las especies de depredadores presentes y relacionar las

tasas de depredación con su estrategia de búsqueda de alimento.

Después de los capítulos anteriores se presenta una discusión general de los

principales resultados obtenidos y, de manera resumida, las conclusiones

generales de la presente tesis doctoral.

Cada uno de los capítulos mencionados anteriormente reproduce

íntegramente un manuscrito publicado o en fase de publicación. En ambos casos,

su estado y la referencia completa (cuando proceda) se ha incluido al comienzo de

cada uno de ellos. Las revistas poseen normas de publicación diferentes, por lo que

se ha mantenido el formato de las referencias, figuras y tablas según las normas

editoriales de cada revista.

REFERENCIAS

Alonso, J.C., Alonso, J.A., Muñoz-Pulido, R. 1994. Mitigation of bird collisions with

transmission lines through groundwire marking. Biological Conservation

67: 129–134.

Barrientos, R., Alonso, J.C., Ponce, C. & Palacín, C. 2011. Meta-analysis of the

effectiveness of marked wire in reducing avian collisions with power lines.

Conservation Biology 25: 893-903.

19

Beaulaurier, D.L. 1981. Mitigation of Bird Collisions with Transmission Lines.

Portland, Oregon, Bonneville Power Administration, U.S. Department of

Energy.

Benton, T.G., Vickery, J.A. & Wilson, J.D. 2003. Farmland biodiversity: is habitat

heterogeneity the key? Trends in Ecology and Evolution 18(4): 182-188.

Bevanger, K. 1999. Estimating bird mortality caused by collision and electrocution

with power lines; a review of methodology. In Birds and power lines.

Collision, electrocution and breeding: 29–56. Ferrer, M. & Janss, G.F.E. (Eds).

Madrid: Quercus.

Clough, Y., Kruess, A., Kleijn, D. & Tscharntke, T. 2005. Spider diversity in cereal

fields: comparing factors at local, landscape and regional scales. Journal of

Biogeography 32: 2007–2014.

Donald, P.F., Green, R.E. & Heath, M.F. 2001. Agricultural intensification and the

collapse of Europe’s farmland bird populations. Proceedings of the Royal

Society of London Serie B 268: 25-29.

Donald, P.F., Sanderson, F.J., Burfield, I.J. & Van Bommel, F.P.J. 2006. Further

evidence of continent-wide impacts of agricultural intensification on

European farmland birds. Agriculture, Ecosystems and Environment 116:

189–196.

European Bird Census council. 2014. Trends of common birds in Europe, 2014

update. http://www.ebcc.info/index.php?ID=557

European commission. Agriculture and rural development. 2011.

http://ec.europa.eu/agriculture

Farago, S. 1986. Investigations on the nesting ecology of the great bustard (Otis t.

tarda L., 1758) in the Devavanya nature conservation district. 1

Comparative studies of microclimate. Aquila 92: 133-173.

Ferrer, M. 2012. Aves y tendidos eléctricos. Del conflicto a la solución. Endesa S.A.y

Fundación Migres, Sevilla, 2012.

20

Fuller, R.J., Norton, L.R., Feber, R.E., Johnson, P.J., Chamberlain, D.E., Joys, A.C.,

Mathews, F., Stuart, R.C., Townsend, M.C., Manley, W.J., Wolfe, M.S.,

Macdonald, D.W. & Firbank, L.G. 2005. Benefits of organic farming to

biodiversity vary among taxa. Biology Letters 1: 431–434.

Gardner, B. 1996. European Agriculture: Policies, Production and Trade, Routledge.

Heijnis, R. 1980. Bird mortality from collision with conductors for maximum

tension. Ecology of Birds 2: 111-129.

Kleijn, D. & Sutherland, W.J. 2003. How effective are European agri-environment

schemes in conserving and promoting biodiversity? Journal of Applied

Ecology 40: 947-969.

Krebs, J.R., Wilson, J.D., Bradbury, R.B. & Siriwardena, G.M. 1999. The second silent

spring? Nature 400: 611-612

Manville, A.M. II. 2009. Towers, turbines, power lines, and buildings: steps being

taken by the U.S. Fish and Wildlife Service to avoid or minimize take of

migratory birds at these structures. Proceedings 4th international Partners

in Flight conference. McAllen: Partners in Flight. pp 262–272.

Martín, B. 2008. Dinámica de población y viabilidad de la Avutarda común en la

Comunidad de Madrid. Tesis doctoral.

Newton, I. 2004. The recent declines of farmland bird populations in Britain: an

appraisal of causal factors and conservation actions. Ibis 146: 579–600.

Palacín, C., Alonso, J.C., Martín, C.A., Alonso, J.A., Magaña, M. & Martín, B. 2004.

Avutarda Común (Otis tarda). In: Madroño, A., González, C., Atienza, J.C.

(eds) Libro Rojo de las Aves de España. Dirección General para la

Biodiversidad-SEO/BirdLife, Madrid, pp 209–213.

Real Decreto 1432/2008 de 12 de septiembre. Ministerio de Economía y Hacienda.

BOE 222, de 13 de septiembre.

Robinson, R. A. & Sutherland, W. J. 2002. Post-war changes in arable farming and

biodiversity in Great Britain. Journal of Applied Ecology 39: 157-176.

21

Sans, F. X., Armengot, L., Bassa, M., Blanco-Moreno, L. J. M., Caballero- López, B.,

Chamorro, L. & José-María, L. 2013. La intensificación agrícola y la

diversidad vegetal en los sistemas cerealistas de secano mediterráneos:

implicaciones para la conservación. Ecosistemas 22(1): 30-35.

Sergio, F., Marchesi, L., Pedrini, P., Ferrer, M. & Penteriani, V. 2004. Electrocution

alters the distribution and density of a top predator, the eagle owl Bubo

bubo. Journal of Applied Ecology 41: 836–845.

Siriwardena, G. M., Baillie, S.R., Buckland, S.T., Fewster, R.M., Marchant, J.H. &

Wilson, J.D. 1998. Trends in the abundance of farmland birds: a quantitative

comparison of smoothed Common Birds Census indices. Journal of Applied

Ecology 35: 24–43.

Suárez, F., Sainz, H., Santos, T. & Gonzalez, F. 1991. Las estepas ibéricas. Centro de

Publicaciones. Secretaría General Técnica. Ministerio de Obras Públicas y

Transportes. Madrid.

Valverde, A. 1958. Aves esteparias de la Península Ibérica. Publicaciones del

Instituto de Biología Aplicada, 27: 41-48.

22

23

CAPÍTULO 1

24

Este capítulo reproduce íntegramente el siguiente artículo:

Ponce, C., Alonso, J.C., Argandoña, G., García Fernández, A. & Carrasco, M. 2010.

Carcass removal by scavengers and search accuracy affect bird mortality estimates

at power lines. Animal Conservation 13(6): 603–12.

25

CAPÍTULO 1

Carcass removal by scavengers and search

accuracy affect bird mortality estimates at

power lines

Carlos Ponce1, Juan Carlos Alonso1, Gonzalo Argandoña1, Alfredo

García Fernández2, Mario Carrasco3

1 Dep. Ecología Evolutiva, Museo Nacional de Ciencias Naturales, CSIC, José

Gutiérrez Abascal 2, E-28006 Madrid, Spain

2 Edificio Departamental I, Campus de Móstoles, Universidad Rey Juan Carlos I,

Madrid, Spain

3 Departamento Medio Ambiente, Prointec, Madrid, Spain

26

ABSTRACT

Bird mortality as a result of collisions with power lines has been of increasing concern in

recent decades, but the real impact on bird populations requires an experimental

assessment of scavenger removal rates and searcher detection errors. Farmland and

steppe birds, two of the most threatened avian groups, have been shown to be particularly

vulnerable to collision with power lines, but few removal and detectability studies have

been developed in cereal farmland habitats, and none in the Mediterranean region. We

conducted five carcass disappearance trials in central Spain by placing 522 corpses of

different sizes under power lines, and searching for remains four times during the

following month. The influence of several factors was examined using multivariate

approach. The accumulated number of carcasses removed by scavengers increased

logarithmically, with 32% removed over the 2-day period after the initial placement, but

only 1.5% removed on a daily basis by day 28. Small birds disappeared earlier and at a

higher proportion than larger birds. Carcass removal rates were site-dependent, but were

not influenced by carcass density or season. The detection rate increased with the

observer’s previous experience and carcass size. Carcass counts at power lines notably

underestimate bird casualties. Our 4-week disappearance equations provide a full range of

scavenging rates and observer efficiency correction factors for a wide range of bird

weights. Fortnightly to monthly search frequencies may be adequate to detect medium- to

large-sized corpses, but are insufficient for smaller birds. Finally, all personnel

participating in carcass searches should be trained previously in this task.

Keywords

Carcass disappearance, collision, electrocution, farmland birds, mortality estimate,

searcher detection rate, steppe birds.

27

INTRODUCTION

Electric power lines are known to be a cause of bird mortality, either through

electrocution or collision with the wires (Bevanger, 1994, 1998; Ferrer & Janss,

1999; Bevanger & Broseth 2001; Erickson et al., 2001; APLIC & USFWS, 2005). This

has generated an increasing concern due to the negative effect it may have on

some species that are particularly vulnerable to these mortality factors (Haas,

1980; Ferrer, de la Riva & Castroviejo, 1991; Alonso, Alonso & Muñoz-Pulido,

1994; Janss, 2000; Janss & Ferrer, 2000). The only efficient way to evaluate the

impact of such mortality is to count dead birds in the power line corridor

(Beaulaurier, 1981; Faanes, 1987; Bevanger ,1999). However, because field

researchers cannot continually monitor the power lines, scavengers can be

expected to find and remove a variable portion of the carcasses between the time

of their deaths and the time the next search is conducted. Also, a number of the

carcasses or their remains will be missed by the observers patrolling the line.

Therefore, the results of carcass searches are affected by two main bias sources,

(a) the rate at which carcasses are removed by scavengers, and (b) the ability of

observers to detect corpses or their remains in the field.

A recent review of birds found poisoned after agricultural pesticide

treatment stated that removal rates may vary widely, altering the mortality

estimates based on carcass searches (Prosser, Nattras & Prosser, 2008). Among

possible factors influencing removal rates are features affecting visibility of

corpses such as their size, colour, or vegetation cover, and local and/or seasonal

changes in scavenger abundance and activity (Heijnis, 1980; Beaulaurier, 1981;

Wobeser & Wobeser, 1992; Bevanger, 1999; Morrison, 2002; Ward et al., 2006). As

for the searcher efficiency, it has also been shown to differ extensively with

vegetation type and size of the bird (Wobeser & Wobeser, 1992; Bevanger, 1999;

Morrison, 2002). Scavenger removal rates and efficiency of field workers should

therefore be known to ensure that these bias sources can be corrected to obtain

accurate estimates of bird mortality rates.

28

The objectives of this study were to (a) determine the carcass removal rate

of power line collision victims and the observers’ search bias by means of a series

of trials simulating collisions of birds with power lines in a farmland area in central

Spain, and (b) examine the influence of various potentially relevant factors such as

study site and season, carcass size and density, and vegetation height and cover,

using a multivariate approach. The aims were to (i) obtain correction factors for

these two bias sources which may be used to improve bird fatality estimates at

power lines –although correction factors obtained in our study should be applied

with caution by researchers working in areas with different habitat

characteristics–, and (ii) suggest acceptable periodicities to conduct future carcass

searches at power lines in farmland habitat. Various studies have carried out

similar carcass removal experiments (Prosser et al., 2008; and references therein),

but few have tried to analyse simultaneously the influence of several factors. Most

of these carcass removal studies have been done to estimate mortality after

pesticide treatment in North America or northern Europe (reviewed in Prosser et

al., 2008), some at wind turbines (reviewed in Morrison, 2002; Siriwardena et al.,

2007), and very few analogous studies have been published in relation with

mortality at power lines (e.g. Bevanger, Bakke & Engen, 1994), although there may

be unpublished reports produced by private companies that are not available. To

our knowledge, this is the first study carried out specifically to assess scavenger

removal rates and search efficiency of birds found dead at power lines in

Mediterranean farmland habitats, using a multivariate approach to deal

simultaneously with several underlying variables.

The two most commonly recognized sources of error affecting bird

mortality estimates at power lines or wind turbines are carcass removal by

scavengers and observer detection error (e.g. Bevanger, 1999; Siriwardena et al.,

2007). A widely used estimator of adjusted bird mortality (MA) is therefore MA =

MU/R x p where MU is the unadjusted mortality expressed as number of fatalities

per km of power line, or wind turbine per year, R is the proportion of carcasses

remaining since the last fatality search, and p is the proportion of carcasses found

by observers searching for dead birds. In the present study we provide a full range

of correction factor values for R and p through a month after fatality occurrence, by

conducting four-week long carcass disappearance trials and developing carcass

29

removal and searcher efficiency equations for four different carcass sizes. From

these equations, correction factors for these two main bias sources can be

calculated for any search periodicity up to one month between consecutive search

surveys, and covering a wide range of bird weights (ca. 50-1 000 g). Other minor

adjustments referring to birds injured by the wires but that die elsewhere

remaining undetected (crippling bias), and natural mortality not caused by

collision with the wires (background mortality) are not quantified because they

are usually assumed to be relatively small.

Farmland areas host many endangered bird species which have suffered

alarming population decreases during the last few decades, due mostly to

agricultural intensification but also due to other human-induced causes (Tucker &

Heath, 1994; Siriwardena et al., 1998; Donald, Green & Heath, 2001; Wretenberg et

al., 2006). Among these, the ever-increasing number of power lines built on

farmland areas, where terrain conditions are more suitable for the installation of

these utility structures, is currently an issue of great concern. Farmland and steppe

species are indeed at present the most threatened bird group, with 83% of the

species subject to unfavourable status (BirdLife International, 2004; Burfield,

2005; Santos & Suárez, 2005). Many of these steppe-birds have significant yet

endangered or declining populations in the Iberian Peninsula (Madroño, González

& Atienza, 2004; Santos & Suárez, 2005), and some of them are particularly

vulnerable to the negative effects of power lines (e.g., common cranes or great

bustards, for which collision with power lines has been identified as the main

cause of adult mortality, Alonso & Alonso, 1999; Janns & Ferrer, 2000; Palacín et

al., 2004).

METHODS

Study area

The study was conducted in five Important Bird Areas in Madrid province, along

with a small area in Guadalajara province, central Spain. In each of these areas we

selected 1-2 km long sectors of power lines covering 14 km of 11 different power

lines in total (Figure 1). The terrain is flat to slightly undulated, with an elevation

30

of 740 + 83 m.a.s.l. It is primarily dedicated to cereal cultivation (mainly wheat

Triticum aestivum and barley Hordeum spp.), with minor fields of legumes (Vicia

spp. and Medicago sativa), olive groves Olea europaea and grapevines Vitis vinifera.

Most cereal is grown in a traditional two-year rotation system that creates a

dynamic mosaic of ploughed, cereal and stubble patches over the region. The

climax vegetation of evergreen oak trees (Quercus ilex) and Retama sp. and Thymus

sp. Scrubland has been generally cleared up to small open-wooded tree plots

interspersed within the dominant farmland. White poplars (Populus alba) are also

found in the IBA, although as in the case of oaks, always as single trees or small

groups.

Figure 1. Location of the study area in Madrid region and number of power lines surveyed (in

brackets). A: Casa de Uceda (1), B: IBA 074 Talamanca-Camarma (5), C: IBA 075 Alcarria de Alcalá

(1), D: IBA 073 Cortados y graveras del Jarama (2), E: IBA 393 Torrejón de Velasco-secanos de

Valdemoro (1), F: IBA 072 Carrizales y sotos de Aranjuez (1).

Cereal fields are harvested during late June to early July. Stubbles and

fallows are also used for sheep grazing. These areas hold populations of threatened

bird species such as great bustard Otis tarda (ca. 1 500 individuals, Alonso et al.,

2003), little bustard Tetrax tetrax (ca. 2 600 individuals, García de la Morena et al.,

2006) pin-tailed and black-bellied sandgrouses Pterocles alchata and P. orientalis

31

(ca. 112 and 100 individuals, respectively, Suárez et al., 2006), and montagu's

harrier Circus pygargus (ca. 100 pairs, Arroyo & García, 2007).

Carcass detection and removal by scavengers

Between November 2007 and August 2008 we carried out five carcass

disappearance trials, respectively in November, December, February, April and

August. Each trial started by placing the bird carcasses on the ground under a

power line (20 and 5 carcasses/ km for November and the rest of the months,

respectively). The line was then surveyed four times through the month following

placement (on days 2, 7, 22 and 28; in December it was not possible to carry out

the survey on day 28 due to unfavourable weather conditions). We searched at

uneven intervals because most of the disappearances are known to occur during

the first days after the casualties (e.g. Balcomb, 1986; Prosser et al., 2008). With

the aid of the GPS we went to each site where we had placed a carcass and looked

for it or its remains, recording any track or trace left by scavengers. On the last

survey day of each trial we removed all carcass remains.

In total, 522 carcasses were placed at 0-20 m from the line beneath the

central conductor wire of the power line to simulate natural collisions (Henderson,

Langston & Clark, 1996; Janss, 2000). One hundred and thirty of these carcasses

were female common pheasants (Phasianus colchicus), 130 red-legged partridges

(Alectoris rufa), 130 common quail (Coturnix coturnix), and 132 halves of common

quail carcasses. We chose these species because they are found in the study area;

pheasants were intended to represent a bird of similar size and plumage to great

bustards, the largest species, while common quail halves should represent small

passerines. Using four size classes (pheasants were Large, partridges were

Medium, quail were Small, and half-quail were very Small) allowed us to explore

the effect of carcass size on removal probability. All carcasses used were from wild

birds hunted and later sold for human consumption, thus they were free from the

smell characteristic of poultry farm birds, which might have influenced the

removal rate by scavengers (Bevanger, 1999). For this reason we preferred wild

common quail halves to any other small farm bird like small chicken or ducks.

Significant weight differences existed between the four size category used (p <

0.001 in all cases; common pheasants: 1008.9 g ( 125), n = 20; red-legged

32

partridges: 406.3 g ( 42.0), n = 25; common quail: 109.5 g ( 14.2), n = 25;

common quail halves: 54.1 g ( 6.3), n = 24). All carcasses were aired in a

ventilated and cold room for 24 h prior to placing it under the power line to

eliminate as much as possible any artificial smell remains but avoiding

decomposition due to temperature, which may reduce the attractiveness to

vertebrate scavengers.

We considered that a carcass had been detected by a scavenger when it had

been moved from the initial location, partially eaten, or completely removed. A

carcass disappeared when the remains found were less than 5 feathers, because a

very low number of feathers found during searches for collision casualties cannot

be interpreted as a collision, as these few feathers could have been lost by a bird

during preening, moulting or fighting (e.g. Bevanger, 1999). We searched for

carcasses up to 30 m away from the initial location to account for possible

dragging of the carcass by scavengers. To look at possible differences in removal

rate due to changes in density of carcasses (see e.g. Linz et al., 1991; Wobeser &

Wobeser, 1992; Ward et al., 2006), we placed them at respectively 50 m- (20

carcasses/ km) and 200 m-intervals (5 carcasses/ km) in two winter trials. As no

differences were found, in all other trials we placed carcases at 200 m-intervals.

The placement order of the four size classes was random. For each carcass placed

we recorded UTM coordinates with GPS (Garmin, ± 3 m error), and vegetation

cover and average height (estimated visually in a circle of 3 m radius around the

carcass). Before placing the carcass we made a cut on its ventral side to simulate

the injury caused by the collision with the cable and to avoid differences respect to

the smallest size (common quail halves).

Carcass detection by observers

We explored the influence of the observer’s experience on carcass detectability

during the first two experiments (141 carcasses). The experience was defined as

the total kilometres surveyed under power lines by each observer before the

present study was carried out. Four observers different from those who had placed

the carcasses surveyed the power lines searching for remains. Each of these

surveys was conducted by two observers, one after another separated by ca. 50 m,

33

walking at a slow, regular pace and parallel to the wires of power lines at a

distance no more than 15 m of the central conductor wire. The visibility was good

along all the power line corridors due to low height of the vegetation, so the

observer could find all the remains to a distance up to 50 m. The first observer

searched for remains without knowing where the carcasses had been placed; the

second walked behind him recording both the remains discovered and those not

found by the first observer.

Statistical analyses

To establish the factors influencing the carcass disappearance rate we used a

generalized linear model with a binary response (carcass or its remains

disappeared vs. present on day 28 after placement). As factors we included each

one of the power lines, month, carcass size, and vegetation cover and height, after

appropriate transformations for vegetation variables -natural logarithm

(height+1) and arcsine (√cover)- to attain equal variance and normality (Sokal &

Rohlf ,1987; Fowler, Cohen & Jarvis, 1998). To explore the relative importance of

each explanatory variable, we used the corrected Akaike’s information criterion

(ΔAICc<2) to select the best models from a set of candidate models with different

combinations of predictor variables (Anderson & Burnham, 1999) and interactions

among them.

Once the relevant factors were identified we performed univariate analyses

to further explore their influence on the carcass disappearance rate. We used some

non parametric tests because we investigated several questions about the different

power lines (11) and months (5) that were considered as independent

experiments. (i) Mann-Whitney U-tests to investigate the importance of the carcass

density, by comparing the number of carcasses disappeared between high density -

November experiment- and low density -all other experiments-; (ii) Chi-squared

tests to search for differences among power lines due to variable carcass density;

(iii) Kruskal-Wallis tests to check for seasonal differences between experiments

carried out on different months; (iv) Kruskal-Wallis and Mann-Whitney U-tests to

explore differences due to carcass size; (v) Chi-squared tests with Yates correction

when necessary (Fowler et al., 1998) to look at removal rate differences between

months or power lines. Finally, to describe the removal rate as a function of

34

carcass size we adjusted a logarithmic equation to disappearance data for each

carcass size.

To investigate the effect of the observer’s experience on carcass

detectability we performed a second generalized linear model with logit link

function and a binary response (carcass or remains found vs. not found), using as

factors the observer, carcass size, vegetation height and vegetation cover. We

applied the same variable transformations and model selection criteria used in the

previous analysis. Also, we carried out univariate analyses to explore (i) whether

large carcasses were detected with higher probability than small ones, and (ii)

differences between observers in their ability to find the remains which could be

attributed to their previous experience. As an estimate of experience we used the

kilometres of power line each observer had patrolled looking for collision

casualties prior to this study. We finally adjusted logarithmic equations to

detectability data for each observer.

RESULTS

Carcass detection and removal by scavengers

On the first survey, two days after leaving the carcasses under the power lines,

67.2% of them had been detected by scavengers, with no differences among bird

sizes (2 = 0.94, d.f. = 3, P < 0.82). Detection rate increased to 93.7% during the

second survey (day 7), with no size differences (2 = 0.12, d.f = 3, P < 0.99), and to

99.8% and 100%, respectively for the third and fourth surveys (days 22 and 28).

The accumulated number of carcasses removed by scavengers increased

from the day they were placed following a logarithmic function (Figure 2). On day

two, 32% of the carcasses had already disappeared. An additional 20% of the

carcasses disappeared between days 2 and 7, a further 16% between days 7 and

22, and only 3% between days 22 and 28. Disappearance rates for each survey

date did not change between experiments carried out on different months (P >

0.08 in all cases). On day 28 after placement of the carcasses under the power

lines, 71.5% of the initial sample had disappeared. This carcass disappearance rate

35

was not influenced by carcass density, either considering all power lines together

in a sample (Z = 1.35, P < 0.18, November vs. all other months; 2 = 0.6625, d.f. = 1,

P < 0.42 between two winter tests -November and February- to control for a

possible seasonal effect), or testing each power line separately (P > 0.18 in all

cases).

Figure 2. Accumulative percent of carcasses disappeared on the different survey dates (= day after

carcasses were placed under the power line). Means and SD are given.

The result of the generalized model showed that carcass disappearance on day 28 was influenced by carcass size (at higher speed for smaller carcasses) and power line, with no significant effects of other variables or interactions among them (Table 1).

Table 1. Results of the generalized linear model for carcass disappearance on the last survey date

(day 28 after placing carcasses).

Variable Partial deviance P

Carcass size 76.43 0.001

Power line 28.17 0.001

Month 4.34 0.226

Month*Carcass size 2.37 0.498

Vegetation height 0.00 0.961

Vegetation cover 0.10 0.749

There were three power lines where disappearance rates differed from the

rest: Belvis–Cobeña and El Casar–La Cueva, where disappearance rate was

36

respectively 23% and 19% lower than average), and Pinto–San Martín de la Vega,

where it was 20% higher. The model was highly significant (2 = 133.016, d.f. = 19,

P < 0.001), explaining 39.5% of the total deviance. Carcass size was included in the

first eight models selected as the best subsets (all eight were highly significant, P <

0.001, Table 2), confirming its higher relevance as compared to power line

(included in models 1-4 and 9-11). Vegetation height and cover appeared

respectively in models 2 and 3, as well as in various successive models, all of them

with ΔAICc > 2 (Table 2).

Table 2. Models selected as best significant subsets by the generalized linear model for carcass

disappearance (see Table 1), ranked according to ΔAICc.

Nº Model AICc ΔAICc wia kb Pc

1 Carcass size-Power line 426.78 0 0.534 13 0.001

2 Carcass size-Power line-Vegetation height 428.88 2.10 0.187 14 0.001

3 Carcass size-Power line-Vegetation cover 428.90 2.12 0.185 14 0.001

4 Carcass size-Power line-Vegetation height-Vegetation cover

430.99 4.21 0.065 15 0.001

5 Carcass size 433.86 7.08 0.015 3 0.001

6 Carcass size-Vegetation height 435.84 9.06 0.006 4 0.001

7 Carcass size-Vegetation cover 435.89 9.11 0.006 4 0.001

8 Carcass size-Vegetation height-Vegetation cover 437.74 10.96 0.002 5 0.001

9 Power line 528.95 102.17 0.000 12 0.027

10 Power line-Vegetation height 531.01 104.23 0.000 13 0.041

11 Power line-Vegetation cover 531.05 104.27 0.000 13 0.041

a model weight b number of parameters c significance of the model

The function describing the disappearance rate through the first month for

each carcass size is shown in Figure 3. On survey date 28, 42.5% of large, 62.1% of

medium-sized, 86.9% of small, and 93% of very small carcasses had disappeared,

with significant differences among these values (H 3, 16 = 13.08, P < 0.005). The

differences were significant between large and medium (Z = -2.31, P < 0.021), and

between medium and small (Z = -2.32, P < 0.020), but not between small and very

small carcasses (Z = -1.15, P < 0.25). Disappearance rates for each carcass size did

not change with carcass density (P > 0.54, P > 0.47, P > 0.46, and P > 0.50,

respectively from large to very small), or power line (2 = 0.28 < P 0.99). Using the

37

weights of the four size classes we obtained an equation predicting the

disappearance rate at 28 days as a function of weight (Figure 4).

Figure 3. Accumulative percent of carcasses of each size disappeared through the four surveys

(days 2, 7, 22 and 28; survey dates were transformed as x=day+1). For each carcass size, five data

corresponding to the five trials conducted on different months are represented (November,

December, February, April and August; in December it was not possible to carry out the survey on

day 28 due to unfavourable weather conditions). The curves represent the logarithmic models that

fitted best to these monthly disappearance figures. Large size: y = 0.744+28.063*log10(x) (r =

0.83); Medium size: y = -1.751+41.880*log10(x) (r = 0.88); Small size: y = -6.623+58.111*log10(x)

(r = 0.84); Very small size: y = 13.538+60.342*log10(x) (r = 0.75). P < 0.001 in all cases.

38

Figure 4. Percent carcasses disappeared on the last survey (day 28) for each bird weight class.

Black dots are the values for the four trials (November, February, April and August). The curve

represents the logarithmic equation adjusted to these values: y = 166.295-40.567*log10(x) (r =

0.93, P < 0.001).

Carcass detection by observers

On average, an observer discovered 53% of the carcasses present. However, there

were significant differences in their ability to find the remains (2 = 3.88, d.f. = 1, P

< 0.05; observers A, B, C and D found respectively 25%, 57.1%, 68.4% and 70.4%

of the carcasses). The generalized model showed that carcass detectability was

influenced by carcass size and observer, with no significant effect of vegetation

height or cover and their interaction (Table 3). The model was highly significant

(2 = 38.56, d.f. = 7, P < 0.001), explaining 20.0% of the total deviance. Large

carcasses were detected in higher proportion (71.7 %) than other sizes

(respectively, 55.8 %, 32.1 %, and 33.3 % for medium-sized, small, and very small

carcasses, 2 = .03, d.f. = 1, P < .05), with no differences among medium to very

small sizes (P > 0.08 in all cases). Fifteen significant candidate models were

obtained, of which the first two showed ΔAICc < 2 and included observer (not in

model 2), carcass size and vegetation height (Table 4). Using the kilometres of

power line surveyed by each observer prior to this study as an index of his

experience in detecting carcasses, this factor explained 92% of the variation of the

detection rate (Figure 5).

39

Table 3. Results of the generalized linear model for carcass detectability.

Variable Partial deviance P

Carcass size 16.42 0.001

Observer 8.38 0.039

Vegetation height 2.26 0.133

Vegetation cover 0.00 0.965

Vegetation height*Vegetation cover 1.87 0.140

Table 4. Models selected as best significant subsets by the generalized linear model for carcass

detection rate (see Table 1), ranked according to ΔAICc.

Nº Model AICc ΔAICc wia kb Pc

1 Observer-Carcass size-Vegetation height 171.18 0 0.431 7 0.001

2 Carcass size-Vegetation height 173.07 1.89 0.166 4 0.001 3 Observer-Carcass size-Vegetation height-Vegetation cover

173.42 2.25 0.140 8 0.001

4 Observer-Carcass size-Vegetation cover 173.43 2.26 0.140 7 0.001

5 Carcass size-Vegetation height-Vegetation cover 175.15 3.98 0.059 5 0.001

6 Observer-Carcass size 176.15 4.98 0.036 6 0.001

7 Carcass size-Vegetation cover 176.84 5.67 0.025 4 0.001

8 Observer-Vegetation height 185.42 14.24 0.000 6 0.001

9 Carcass size 185.75 14.58 0.000 3 0.001

10 Observer-Vegetation cover 186.24 15.07 0.000 6 0.001

11 Observer 186.99 15.82 0.000 5 0.001

12 Observer-Vegetation height-Vegetation cover 187.60 16.42 0.000 7 0.001

13 Vegetation height 188.60 17.42 0.000 3 0.001

14 Vegetation height-Vegetation cover 190.58 19.41 0.000 4 0.006

15 Vegetation cover 190.64 19.47 0.000 3 0.004

a model weight b number of parameters c significance of the model

40

Figure 5. Detection ability of the four observers participating in the detectability trial (black dots),

as a function of their experience (defined as the number of kilometres of power line surveyed prior

to the present study). The curve represents the equation adjusted to the four detection ability

values, y = 24.461+13.827*log10(x) (r = 0.961, P < 0.04).

DISCUSSION

Our results indicate that removal of carcasses by scavengers reduced the number

of dead birds placed initially under power lines. The number of carcasses present

followed a logarithmically decreasing trend through the days following trial start.

Second, searcher efficiency biased the number of carcasses low to a lower level by

a variable extent, depending on previous personal training. Third, these two

sources of bias increased with decreasing carcass size, and removal rate was also

site-dependent. Fourth, the corresponding corrections should be taken into

account when using carcass surveys to calculate bird mortality estimates due to

electrocution or collision at power lines. Below we discuss these results in detail.

These conclusions can be drawn from our study, in spite of the following

methodological limitations which could have affected the scavenging rates

obtained. For example, our presence in the area and handling of the carcasses

when placing them may have either attracted or deterred scavengers. Scavengers

could have followed human trails to carcasses or, alternatively, shy species might

have avoided carcasses or sites tainted with human scent (Wobeser & Wobeser,

41

1992). We believe, however, that these effects were negligible because in our study

area scavengers are likely to be used to human presence due to the frequent

occurrence of human activities such as farming, sheepherding and hunting. We

tried to minimize other possible sources of error based on carcass odour or

conspicuousness. The results of previous studies have suggested that brighter-

coloured corpses may be more conspicuous and easier to be detected by aerial

scavengers (e.g. Balcomb, 1986; Prosser et al., 2008). This would however not

influence removal rates by mammalian scavengers, which mostly search by scent

and are nocturnal. More frequently, authors have drawn attention to the removal

rates between wild bird carcasses and those of artificially reared species (Balcomb,

1986; Young et al., 2003; Prosser et al., 2008). We used exclusively wild birds shot

by hunters to minimize these odour-based effects. Moreover, we left corpses one

day aired in a ventilated and cold room before placing them to eliminate any scent

from handling by hunters and suppliers. Also, the species we used belonged to the

local fauna and were similar in plumage colour and pattern to most other steppe-

birds living in the study area. Another source of variation in removal rate may be

the carcass density. Obviously, in carcass removal trials carcass density is higher

than in most natural events, in order to make searches and calculations feasible

within reasonable time and space limits. Some authors have suggested that greater

than normal carcass abundance may attract scavengers and either increase

removal rate (Bevanger et al., 1994), decrease it due to satiation (Linz et al., 1991),

or produce no observable effects (Wobeser & Wobeser, 1992; Prosser et al. 2008).

Another studies carried out in the same power lines by the authors showed that

around 8 wild dead birds/ km were found under them during one year sampling

(one each month) so, if we consider that many of the collided or electrocuted birds

may have been moved by scavengers or not found by the observers (as we have

demonstrated in the present work), we can assume that we have not significantly

increased the density of dead birds with respect to normal casualties. But to check

for this possibility in this experiment, we compared our standard density with a

four-fold density, and found no differences in removal rate.

Carcasses were removed by scavengers with highest intensity immediately

after placement. Later, removal rate decreased regularly through a period varying

between some days and several weeks. The accumulative disappearance curves

42

best fitting the data were logarithmic and similar in shape for all four size classes

tested, but smaller carcasses disappeared earlier and in higher proportion than

larger ones. Our results show that removal rate increased with decreasing carcass

size, except for the two smallest size classes which were removed at similar rates.

These smaller carcasses were most frequently removed without leaving any

remains (66.7% small and 85.7% very small carcasses removed on day 2), in

contrast to big corpses which were normally partially eaten on the spot (on day 2,

78.8% medium and 73.6% large corpses; all size differences significant, P < 0.02).

Remains of larger corpses were easily recognized through the whole series of

search surveys, most often ending up as a pile of feathers that usually remained for

several weeks on the spot, indicating past scavenger activity. These facts suggest

that a wider spectrum of scavenger species were able to feed on and remove

corpses below a certain size, whereas potential predators able to remove larger

carcasses at once were much scarcer, and these large corpses were discovered and

as a rule incompletely devoured by the same scavengers as those feeding on the

smaller corpses. Common scavengers in our study area include mammals like fox

(Vulpes vulpes), feral dogs (Canis familiaris), feral cats (Felis silvestris catus) or

black rat (Rattus rattus), and birds such as black and red kites (Milvus migrans and

M. milvus), corvids like magpies (Pica pica), jackdaws (Corvus monedula), ravens (C.

corax), white storks (Ciconia ciconia), and black-headed and black-backed Gulls

(Larus ridibundus and L. fuscus). The fact that we didn’t find differences among

carcass sizes in the scavenger detection rate (which includes both disappeared and

partially eaten carcasses) indicates that corpses were found opportunistically, and

not due to their visibility. This suggests that the most frequent scavengers in our

study area were probably mammals, which mostly hunt by scent (see also

Kostecke, Linz & Bleier, 2001 for the same interpretation based on results

confirmed through photographic evidence). Smallwood et al., (2008) found 74%

and 63% of the carcasses respectively detected and removed by mammals,

although in that study differences among carcass sizes were found. However,

identification of all scavenger species and their relative contribution to the

disappearance of carcasses was not among our objectives.

Previous studies have also found decelerating removal rates (Balcomb,

1986; Ward et al., 2006), and very high initial removal rates among smaller

43

carcasses, most of which disappeared within the first days (Heijnis, 1980; Wobeser

& Wobeser, 1992; Prosser et al., 2008). However, few of these studies followed

carcasses for more than a week, which renders estimates of the eventual fate of

certain carcasses difficult, particularly of the larger ones which usually survive

longer. In our study we surveyed the power lines through four weeks after

placement, because one of our main objectives was to determine the frequency

with which carcass searches should be conducted to determine fatalities at power

lines. Although most mortality studies at power lines are based on weekly to

monthly survey frequencies, such periodicity is usually fixed without a well-

founded basis. The disappearance curves obtained in our study through a month

for various bird sizes offers the opportunity to determine an acceptable search

frequency, depending on the bird species for which removal rates are required. An

interesting result not found in most previous studies was that for all four carcass

sizes tested, further removals were recorded even over 20 days after placing the

corpses.

The second factor influencing removal rate was the power line. No

significant effects were found from other variables such as season or vegetation

structure, suggesting a relatively uniform scavenger pressure through the year and

among different substrate types. Changes in scavenger density have been

suggested to be the main reason for the differences in removal rate found among

sites (Kostecke et al., 2001), seasons (Bevanger et al., 1994; Linz et al., 1991;

Johnson et al., 2003; Prosser et al., 2008), or areas with different vegetation

structure (Bevanger et al., 1994; Bevanger, 1995; Siriwardena et al., 2007). In our

study, only three power lines showed unusual removal rates. The line Pinto - San

Martín de la Vega was close to a huge rubbish dump, where large numbers of black

kites and white storks are found in spring and summer, and black-headed and

lesser black-backed gulls aggregate by thousands, mainly in winter. Individuals of

all these species have wide home ranges and could have easily contributed to the

higher carcass removal rate recorded at this power line. The two power lines with

lowest removal rates were located in close proximity to villages, which might have

determined a lower density of scavengers and, therefore, a lower removal rate.

However, the purpose of our study was only to explore the relative amount of local

or seasonal differences and their effect on removal rate, nor to investigate the

44

causes of such differences. Based on the significant differences found in three of

eleven lines, we conclude that scavenger rates are probably site-dependent in most

cases. Moreover, although seasonal differences in removal rate did not reach

statistical significance in our study, the range of values obtained for different

months was quite wide, which suggests that seasonal variation could be an

important factor to be considered in future studies. A similar conclusion can be

drawn for vegetation structure, which did not appear to significantly affect

removal rate, but appeared on some of the candidate models selected in our

analyses. Overall, this suggests that local, seasonal, and other differences due to

vegetation structure, may affect scavenger removal rate to a variable extent, and

therefore the figures given in the present study should be taken with care. For

example, a more dense, diverse or higher vegetation could be an influential

variable in studies focusing on small birds. The correction indices derived from our

trials could probably be applied to estimate power line-caused mortality in similar

habitats within the Mediterranean region, being less useful for areas differing

much in geographic location, habitat structure or scavenger community. Studies

similar to the present one should be conducted in areas with completely different

climatic conditions, i.e. where the ground is covered with snow through several

months in winter, or the vegetation and habitat structure are quite different, in

order to check the importance of weather and vegetation variables and obtain

more reliable correction factors.

Finally, the four observers participating in this study differed notably in

their ability to find carcasses (25-70.4%). A similar range in detectability values

has also been reported in previous studies (e.g., 35-85% in Morrison’s 2002

review). Lower detection rates have been attributed to a higher (Philibert,

Wobeser & Clark, 1993) or denser (Wobeser & Wobeser, 1992) vegetation. In our

farmland study area, changes in vegetation structure were probably not enough to

determine significant variations in detectability. The two factors that we found to

influence detectability were carcass size and previous experience of the observer.

Larger carcasses were detected in higher proportion than smaller ones, as

reported in Siriwardena’s (2007) review of wind turbine-caused mortality. The

correlation found in our study between detection rate and previous experience of

the observer specifically conducting these kinds of searches at power lines is an

45

important new result that highlights the importance of a training period for field

workers participating in carcass searches intended to estimate mortality rates at

power lines. We cannot exclude that other factors, e.g. personal motivation may

influence the searching detection rate. Finally, the results should be interpreted

with caution, due to the small number of observers that have participated in the

experiment.

CONCLUSIONS AND MANAGEMENT IMPLICATIONS

Carcass counts at power lines will notably underestimate the number of bird

casualties, the bias being higher in smaller birds. Mortality estimates should

incorporate correction factors based on scavenging rates and observer efficiency.

Conservation authorities and power line operators should be aware of these bias

sources and adjust past and future estimates before using them to assess power

line-caused bird mortality. Scavenger removal rates differed to a great extent with

carcass size, being much higher for small birds. A high percent of these small

carcasses had disappeared two days after placement, and ca. 90% after two weeks.

This indicates that fortnightly to monthly search frequencies may be adequate to

detect casualties of medium to large-sized species, but are insufficient in the case

of smaller species. For these, a higher search frequency is recommended, in order

to reduce the uncertainty interval implicit in extrapolations from equations like

those presented here. Although site-related and seasonal differences found in our

study did not reach statistical significance, the range of values obtained for a

sample of 55 surveys (5 months × 11 power lines) was considerable. This suggests

that, if precise mortality estimates are required, scavenger removal trials should

be carried out simultaneously with searches aiming to estimate collision mortality.

We recommend carrying out such complementary removal trials whenever

possible. Alternatively, the equations presented here may be used to obtain

mortality estimates in Mediterranean farmland. Figures may vary substantially

between this and other farmland habitats at different latitudes. Therefore, similar

studies are needed in these habitats to evaluate the effects of various bias sources

affecting scavenger removal rates there. Finally, all personnel participating in

46

carcass searches should be previously trained in this task, in order to minimize

detection errors due to low experience.

ACKNOWLEDGEMENTS

We thank C. Bravo for help during field work and L.M. Carrascal for help during

statistical analysis. P. Prosser and K.S. Smallwood reviewed an early version of the

manuscript. Two anonymous referees and A. Amar (Associate Editor of Animal

Conservation) improved the manuscript with their comments. Funds were

provided by a contract CSIC-HENARSA to set up and evaluate steppe-bird

conservation measures at IBA 074, and by project CGL2008-02567/BOS of the

Dirección General de investigación.

REFERENCES

Alonso, J.A. & Alonso, J.C. (1999). Mitigation of bird collisions with transmission

lines through groundwire marking. In Birds and Power Lines. Collision,

Electrocution and Breeding: 113-121. Ferrer, M. & Janss, G.F.E. (Eds).

Madrid: Quercus.

Alonso, J.C., Alonso, J.A. & Muñoz-Pulido, R. (1994). Mitigation of bird collisions

with transmission lines through groundwire marking. Biol. Conserv. 67, 129-

134.

Alonso, J. C., Martín, C.A., Palacín, C., Magaña, M. & Martín, B. (2003). Distribution,

size and recent trends of the Great Bustard (Otis tarda) population in

Madrid region, Spain. Ardeola 50, 21-29.

Anderson, D.R. & Burnham, K.P. (1999). General strategies for the analysis of

ringing data. Bird Study 46, 261-270.

Arroyo, B. & García, J. (2007). El aguilucho cenizo y el aguilucho pálido en España.

Población en 2006 y método de censo. Madrid: SEO/BirdLife.

47

Avian Power Line Interaction Committee (APLIC) & U.S. Fish and Wildlife Service

(USFWS). (2005) Avian Protection Plan (APP) Guidelines. Washington, D.C:

Edison Electric Institute/U.S. Fish & Wildlife Service.

Balcomb, R. (1986). Songbird carcasses disappear rapidly from agricultural fields.

Auk 103, 817-820.

Beaulaurier, D.L. (1981). Mitigation of bird collisions with transmission lines.

Oregon: Bonneville Power Administration Report. U. S. Dept. of Energy.

Bevanger, K. (1994). Bird interactions with utility structures; collision and

electrocution, causes and mitigating measures. Ibis 136, 412-425.

Bevanger, K. (1995). Estimates and population consequences of tetraonid

mortality caused by collisions with high tension power lines in Norway. J

Appl. Ecol. 32, 745-753.

Bevanger, K. (1998). Biological and conservation aspects of bird mortality by

electricity power lines: a review. Biol. Conserv. 86, 67-76.

Bevanger, K. (1999). Estimating bird mortality caused by collision and

electrocution with power lines; a review of methodology. In Birds and

Power Lines. Collision, Electrocution and Breeding: 29-56. Ferrer, M. & Janss,

G.F.E. (Eds). Madrid: Quercus.

Bevanger, K., Bakke, O. & Engen, S. (1994). Corpse removal experiment with willow

ptarmigan (Lagopus lagopus) in power-line corridors. Ökol. Vögel (Ecol.

Birds) 16, 597-607.

Bevanger, K. & Broseth, H. (2001). Bird collisions with power lines -an experiment

with ptarmigan (Lagopus spp.). Biol. Conserv. 99, 341-346.

Bird/Life International. (2004). Birds in Europe. Population estimates, trends and

conservation status. Bird/Life Conservation series nº 12. Cambridge:

Bird/Life International.

Burfield, I. (2005). The conservation status of steppic birds in Europe. In Ecology

and Conservation of Steppe-land Birds: 119-131 Bota, G., Morales, M.,

48

Mañosa, S. & Camprodon, J. (Eds.). Barcelona: Lynx Edicions & Centre

Tecnològic Forestal de Catalunya.

Donald, P.F., Green, R.E. & Heath, M.F. (2001). Agricultural intensification and the

collapse of Europe’s farmland bird populations. Proc. R. Soc. B. 268, 25-29.

Erickson, W.P., Johnson, G.D., Strickland, M.D., Young, D., Sernka, K.J. & Good, R.E.

(2001). Avian collisions with wind turbines: a summary of existing studies and

comparison to other sources of avian collision mortality in the United States.

Washington, D. C: Western EcoSystems Technology Inc. National Wind

Coordinating Committee (NWCC).

Faanes, C.A. (1987). Bird behaviour and mortality in relation to power lines in

prairie habitats. Washington, D.C: U. S. Fish Wildlife Service Technical,

report 7.

Ferrer, M., de la Riva, M. & Castroviejo, J. (1991). Electrocution of raptors on power

lines in southwestern Spain. J. Field Ornithol. 62, 181-190.

Ferrer, M. & Janss, G.F.E. (1999). Birds and Power Lines. Collision, Electrocution and

Breeding. Madrid: Quercus,

Fowler, J., Cohen, L. & Jarvis, P. (1998). Practical statistics for field biology, 2nd edn.

Chichester: John Wiley & Sons.

García de la Morena, E.L., Bota, G., Ponjoan, A. & Morales, M. B. (2006). El sisón

común en España. I Censo Nacional (2005). Madrid: SEO/BirdLife.

Haas, D. (1980). Gefährdung unserer GroBvögel durch Stromschlag eine

Dokumentation. Ökol. Vögel (Ecol. Birds) 2, 7-57.

Heijnis, R. (1980). Vogeltod durch Drahtanflüge bei Hochspannungsleitungen. Ökol.

Vögel (Ecol. Birds) 2, 111-129.

Henderson, I.C., Langston, R.H.W. & Clark, N.A. (1996). The response of common

terns Sterna hirundo to power lines: an assessment of risk in relation to

breeding commitment, age and wind speed. Biol. Conserv. 77, 185-192.

49

Janss, G.F.E. (2000). Avian mortality from power lines: a morphological approach

of a species-specific mortality. Biol. Conserv. 95, 353-359.

Janss, G.F.E. & Ferrer, M. (2000). Common crane and great bustard collision with

power lines: collision rate and risk exposure. Wildlife Soc. B. 28, 675-680.

Johnson, G., Erickson, W., White, J. & McKinney, R. (2003). Avian and Bat Mortality

During the First Year of Operation at the Klondike Phase I Wind Project,

Sherman County, Oregon. Cheyenne: Western Ecosystems Technology Inc.

Kostecke, R. M., Linz, G. M. & Bleier, W. J. (2001). Survival of avian carcasses and

photographic evidence of predators and scavengers. J. Field Ornithol. 72,

439-447.

Linz, G., Davis, J.E. JR., Engeman, R.M., Otis, D.L. & Avery, M.L. (1991). Estimating

survival of bird carcasses in cattail marshes. Wildlife Soc. B. 19, 195-199.

Madroño, A., González, C. & Atienza, J.C. (2004). Libro Rojo de las Aves de España.

Madrid: Dirección General para la Biodiversidad & SEO/BirdLife.

Morrison, M.L. (2002). Searcher bias and scavenging rates in bird-wind energy

studies. Colorado: National Renewable Energy Laboratory.

Palacín, C. Alonso, J.C., Martín, C.A., Alonso, J.A., Magaña, M. & Martín, B. (2004).

Avutarda Común. In Libro Rojo de las Aves de España: 209-213. Madroño, A.,

González, C. & Atienza, J.C. (Eds.). Madrid: Ministerio de Medio Ambiente &

SEO/BirdLife.

Philibert, H., Wobeser, G. & Clark, R. G. (1993). Counting dead birds: Examination of

methods. J. Wildlife Dis. 29, 284-289.

Prosser, P., Nattras, C. & Prosser, C. (2008). Rate of removal of bird carcasses in

arable farmland by predators and scavengers. Ecotox. Environ. Safe. 71, 601-

608.

Santos, T. & Suárez, F. (2005). Biogeography and population trends of iberian

steppe birds. In Ecology and Conservation of Steppe-land Birds: 69-102 Bota,

50

G., Morales, M., Mañosa, S. & Camprodon, J. (Eds.). Barcelona: Lynx Edicions

& Centre Tecnològic Forestal de Catalunya.

Siriwardena, G.M., Baillie, S.R., Buckland, S.T., Fewster, R.M., Marchant, J.H. &

Wilson, J.D. (1998). Trends in the abundance of farmland birds: a

quantitative comparison of smoothed Common Birds Census indices. J. Appl.

Ecol. 35, 24-43.

Siriwardena, G.M., Baillie, S.R., Buckland, S.T., Fewster, R.M., Marchant, J.H. &

Smallwood, K.S. (2007). Estimating wind turbine-caused bird mortality. J.

Wildlife Manage. (Suppl.) 71, 2781-2791.

Smallwood, K.S., Lee, A.N., Douglas, A.B., Joseph, E.D., Brian, R., Karas, S.A.S. &

Salvador, R.L. (2008). Range Management Practices to Reduce Wind Turbine

Impacts on Burrowing Owls and Other Raptors in the Altamont Pass Wind

Resoure Area, California. California: California Energy Commission, PIER

Energy‐Related Environmental Research Program. CEC‐500‐2008‐080.

Sokal, R.R. & Rohlf, F.J. (1987). Introduction to biostatistics. San Francisco: W.H.

Freeman & Company.

Suárez, F., Hervás, I., Herranz, J. & Del Moral, J.C. (2006). La ganga ibérica y la

ganga ortega en España: población en 2005 y método de censo. Madrid:

SEO/BirdLife.

Tucker, G.M. & Heath, M.F. (1994). Birds in Europe: Their Conservation Status.

Cambridge: Bird/life International.

Ward, M.R., Stallknecht, D.E., Willis, J., Conroy, M.J. & Davidson, W.R. (2006). Wild

bird mortality and West Nile virus surveillance: Biases associated with

detection, reporting and carcass persistence. J. Wildlife Dis. 42, 92-106.

Wobeser, G.A. & Wobeser, A.G. (1992). Carcass disappearance and estimation of

mortality in a simulated die-off of small birds. J. Wildlife Dis. 28, 548-554.

Wretenberg, J., Lindström, Å., Svensson, S., Tierfelder, T. & Pärt, T. (2006).

Population trends of farmland birds in Sweden and England: similar trends

51

but different patterns of agricultural intensification. J. Appl. Ecol. 43, 1110-

1120.

Young, D.P. Jr., Erickson, W.P., Good, R.E., Strickland, M.D. & Johnson, G.D. (2003).

Avian and Bat Mortality Associated with the Initial Phase of the Foote Creek

Rim Windpower Project, Carbon County, Wyoming: November 1998-June

2002. Wyoming: Western EcoSystems Technology, Inc.

52

53

CAPÍTULO 2

54

Este capítulo reproduce íntegramente el siguiente artículo:

Barrientos, R., Ponce, C., Palacín, C., Martín, C.A., Martín, B. & Alonso, J.C. 2012. Wire

marking results in a small but significant reduction in avian mortality at power

lines: A BACI designed study. Plos One 7(3): e32569.

doi:10.1371/journal.pone.0032569.

55

CAPÍTULO 2

Wire marking results in a small but

significant reduction in avian mortality at

power lines: A BACI designed study

Rafael Barrientos1 *, Carlos Ponce1, Carlos Palacín1, Carlos A.

Martín1 2, Beatriz Martín1 3 & Juan Carlos Alonso1

1 Departamento de Ecología Evolutiva, Museo Nacional de Ciencias Naturales

(CSIC), Madrid, Spain. Tel: +34 91 411 13 28; Fax: +34 91 564 50 78

* Current address: Área de Zoología, Departamento de Ciencias Ambientales,

Facultad de Ciencias del Medio Ambiente, Universidad de Castilla-La Mancha,

Toledo, Spain.

2 Current address: Departamento de Zoología y Antropología Física, Facultad de

Biología, Universidad Complutense de Madrid, Madrid, Spain.

3 Current address: Fundación Migres, ,Huerta Grande, Pelayo, Algeciras, Cádiz,

Spain.

56

ABSTRACT

Background: Collision with electric power lines is a conservation problem for many bird

species. Although the implementation of flight diverters is rapidly increasing, few well-

designed studies supporting the effectiveness of this costly conservation measure have

been published.

Methodology/Principal Findings: We provide information on the largest worldwide

marking experiment to date, including carcass searches at 35 (15 experimental, 20

control) power lines totalling 72.5 km, at both transmission (220 kV) and distribution (15

kV-45 kV) lines. We found carcasses of 45 species, 19 of conservation concern. Numbers of

carcasses found were corrected to account for carcass losses due to removal by scavengers

or being overlooked by researchers, resulting in an estimated collision rate of 8.2

collisions per km per month. We observed a small (9.6%) but significant decrease in the

number of casualties after line marking compared to before line marking in experimental

lines. This was not observed in control lines. We found no influence of either marker size

(large vs. small spirals, sample of distribution lines only) or power line type (transmission

vs. distribution, sample of large spirals only) on the collision rate when we analyzed all

species together. However, great bustard mortality was slightly lower when lines were

marked with large spirals and in transmission lines after marking.

Conclusions: Our results confirm the overall effectiveness of wire marking as a way to

reduce, but not eliminate, bird collisions with power lines. If raw field data are not

corrected by carcass losses due to scavengers and missed observations, findings may be

biased. The high cost of this conservation measure suggests a need for more studies to

improve its application, including wire marking with non-visual devices. Our findings

suggest that different species may respond differently to marking, implying that species-

specific patterns should be explored, at least for species of conservation concern.

57

INTRODUCTION

Bird collisions with electric power lines have raised conservation concerns since

the early 1900s, but it was not until the 1970s that biologists and engineers began

to realize the extent of this problem [1,2]. Today the number of power lines is

increasing worldwide at an annual rate of approximately 5% [3]. Mortality from

collisions with power lines and other electric utility structures has been

documented for some 350 bird species [4]. However, until a cumulative impacts

assessment of power line mortality is conducted, the real level of mortality will

remain uncertain [5]. Only some crude estimates of the importance of the problem,

all of them based on extrapolations, are available. For example, in the Netherlands

it has been found that bird collisions with power lines may cause one million

deaths per year [6]. In the United States, [5] it is estimated that power lines may

kill up to 175 million birds annually, and it is estimated that bird collisions with

power structures, including transmission (≥70 kV, usually with ground-wire and

wires at more than one height) and distribution (<70 kV, commonly without

ground-wire and all the wires at the same height) lines, could approach one billion

avian fatalities per year worldwide [7]. Fortunately, these values are probably

overestimated since most of the studies are usually carried out on power lines that

cause an important number of fatalities. Nevertheless, these figures allow

conservationists to speculate that mortality due to collisions with power lines

represents a serious threat for population viability in many species, at least in

those that undergo higher collision risks, and that this threat is not equal for all

species. Indeed, birds with low manoeuvrability, i.e., those with high wing loading

and low aspect, such as bustards, pelicans, waterfowl, cranes, storks, and grouse,

are among the species most likely to collide with power lines [2,8]. Species with

narrow visual fields are also at high collision risk as they do not see the wires

[9,10]. Despite this potentially important conservation problem, few studies have

analyzed in detail how these losses affect population trends. For instance, it has

been estimated that collision-related losses might equal up to 90% of the annual

number of grouse harvested by hunting in Norway [11]. Based on ring-recovery

data [12], it has been assessed that 25% of juveniles and 6% of adult white storks

58

(Ciconia ciconia) die annually in Switzerland due to power lines (although these

data also include electrocutions). It has also been estimated that 30% of Denham’s

bustards (Neotis denhami) die annually by collisions with power lines in South

Africa [13].

Researchers and managers have used several methods to reduce collisions,

including the removal of the static wire [14,15]. However, the most popular

measure has been the attachment of spirals, plates, swivels, or spheres

(collectively known as bird flight diverters) to the static wire in order to increase

visibility [3,16,17,18]. While a recent review concluded that marking static wires

reduces the overall number of bird casualties at power lines, it also called attention

to the fact that there are a surprisingly small number of well-designed, peer-

reviewed studies to support this [19]. Furthermore, there remain many gaps in the

research in this area, with several important details still unresolved; for example,

the comparative effectiveness of various currently available marker types [19]. To

confirm diverter effectiveness, and to study all details of this conservation measure

in depth is especially important because despite the high costs of wire marking

(e.g., 1,100-2,600 US$ per marked kilometre in South Africa, [20]; 6,000€ in Spain;

[21]), the application of this conservation measure is rapidly increasing

worldwide.

As stated above, it has been shown that the presence of flight diverters was

associated with a decrease in bird collisions [19]. However, the large differences in

wire-marking techniques constrained the ability to evaluate potential differences

among methods (e.g., different performance based on diverter traits) in that

review. To complement such an approach, in the present study we designed the

largest field experiment to date, to investigate: (i) the effectiveness of wire

marking in reducing collisions; and the roles of (ii) power line type (transmission

vs. distribution), and (iii) spiral size on marking effectiveness. We expected that: (i)

the attachment of spirals would reduce bird mortality [19]; (ii) the effectiveness of

marking would be higher in transmission lines because power line type influences

the frequency of reactions to marked spans [22]. Morkill & Anderson [22] found

that whooping cranes (Grus americana) reacted more than expected to

transmission lines (345 kV, 27 m high) whereas the opposite was true in

59

distribution lines (69 kV, 12 m high). It is worth noting that transmission lines in

our study accumulate a larger number of collisions of those groups of birds

especially prone to collision, such as bustards, storks or waterfowl (see below)

compared to distribution lines. Therefore, the improvement margin once spirals

are attached is greater in transmission lines; and, (iii) larger spirals may be more

effective in increasing the visibility of wires [23,24], reducing collisions to a larger

extent.

METHODS

Study area

The study was conducted in five important bird areas (IBAs) in central Spain (see

[25] for details), which are also the main dry cereal farmland areas in the Madrid

region. The terrain is flat to slightly undulating, with a mean elevation of c. 750 m

a.s.l. These areas are primarily dedicated to cereal cultivation (mainly wheat

Triticum aestivum and barley Hordeum spp.), with minor fields of legumes Vicia

spp., grapevines Vitis vinifera and olive Olea europaea groves. Most cereal is grown

in a traditional 2-year rotation system that creates a dynamic mosaic of ploughed,

cereal and stubble patches over the region. Small patches of natural vegetation

(holm oaks Quercus ilex, and scrubland of Retama spp. and Thymus spp.) remain

dispersed across the cereal matrix. Cereal fields are harvested in late June to early

July. Stubbles and fallows are also used for sheep grazing [26].

Study species

We considered all birds that we found dead under the power lines in the study

area. We discarded the dead birds found beside poles whose cause of death could

be attributed to electrocution. However, since not all species have the same

collision risk [2,8,9], it is worth noting that the study area holds significant

populations of threatened species which are prone to high collision rates due to

their low manoeuvrability, high speed flight and/or poor vision [2,8,9], such as the

great bustard Otis tarda (c. 1500 individuals; [27]), little bustard Tetrax tetrax (c.

60

2600 individuals; [28]), pin-tailed sandgrouse Pterocles alchata and black-bellied

sandgrouse P. orientalis (c. 150 and 200 individuals, respectively, [29]).

Study design and power line monitoring

The study was carried out using a before-after-control-impact (BACI) design, i.e.

monitoring power lines before and after the placement of spirals, combined with

the use of controls during similar time intervals. Between August 2001 and

December 2010 we surveyed bird collisions monthly at 22 different power lines, 7

of them transmission (220 kV) and 15 distribution (15 kV-45 kV) lines, totalling

16.1 and 27.0 km, respectively (Table 1). Fifteen of these lines were our

experimental lines, i.e. to which spirals were attached. These were monitored once

per month for two complete years (one year before and one year after wire

marking). Another 7 lines to which no spirals were attached were used as control

lines and were monitored also once per month for two complete years. Because no

more non-marked control lines were available, in addition to these 7 control lines

we also used as controls the second of 10 two-year and the third of 3 three-year

surveys carried out at experimental lines once spirals were attached to them

(Table 1). These surveys can be considered as controls since once the line was

marked no changes occurred in the factor presence/absence of spirals and thus no

changes were expected between years in the variable under study, i.e. collision

rate. The resulting number of power lines (35) and the total length surveyed

monthly (72.5 km) for all study years make our study both the most detailed and

that with the largest number of power lines monitored to date (for instance, the

mean number of power lines per study was 1.9 in a recent review, see Appendix S2

in [19]).

61

Table 1. Power line name, type of line (transmission or distribution), design (experimental or

control) and number of years monitored after spiral attachment.

Power line Type Length (km) Design Times after

Aranjuez E-O Distribution 2.0 Control One

Aranjuez N-S I Transmission 2.0 Experimental One

Aranjuez N-S II Transmission 4.1 Experimental One

Belvis-Cobeña Transmission 3.0 Experimental Three

Camarma-Fresno Distribution 2.0 Experimental Two

Camarma-Meco Transmission 1.6 Experimental Two

Camarma-Torote Transmission 2.1 Experimental Three

Campo Real-Valdilecha Distribution 3.2 Experimental Two

Daganzo-Alcalá Distribution 0.9 Control One

Daganzo-Fresno Rio Distribution 1.1 Control One

Daganzo-Torote Transmission 1.8 Experimental Three

El Colegio Distribution 3.0 Experimental Two

La Cueva-El Casar Distribution 1.5 Control One

Mesones Distribution 2.0 Control One

Pinto Transmission 1.5 Experimental Two

Pozuelo-Valdilecha Distribution 2.6 Experimental Two

Quer Distribution 1.4 Experimental One

San Martín de la Vega Distribution 1.7 Experimental Two

Valdepiélagos-Talamanca I Distribution 2.2 Experimental One

Valdepiélagos-Talamanca II Distribution 0.5 Control One

Valdetorres-La Jara Distribution 1.4 Control One

Villanueva-Quer Distribution 1.5 Experimental One

One month before the beginning of each monitoring year we removed all

carcasses under the power line. Each monthly search for bird carcasses was

carried out by one observer walking at a slow, regular pace parallel to the wires

but making zigzags to reasonably visually cover a 25 m band at each side of the

vertical of the central conductor wire. The observer surveyed first one side along

the line (e.g. the 25 m band on the right side), and then he/she returned to the

starting point surveying the other side (25 m band on the left side). All remains

found were identified to the species level and removed to avoid double counts.

When the species was unknown (<2% of the cases), the carcass was assigned to

one of the four sizes considered (see below). We recorded a carcass when the

remains found consisted of more than five feathers in a square meter, because a

smaller number of feathers cannot safely be interpreted as a collision, since they

62

could have been lost by a bird during preening, moulting or fighting [30]. Carcass

searches were not performed in June because crop height may lead to

unrealistically low carcass detection figures. July surveys were always carried out

after cereal harvesting. However, it is worth noting that in our rather structurally-

homogeneous study area, there was no relationship between vegetation height or

cover and carcass detection rates [25].

Potential detection biases such as site- or year-dependent carcass removal

by scavengers or variation in carcass detection due to habitat heterogeneity are

minimized in our study, since we used a BACI design combined with the use of

control power lines at the same time intervals. Furthermore, potential outbreaks in

scavenger populations are unexpected because predator control is widespread in

our study region [31]. However, since monthly search frequencies may be

adequate to detect medium- to large-sized corpses, but are insufficient for smaller

birds, we used equations from [25] to adjust our mortality estimates in relation to

search periodicity and carcass size (Table 2), because both can influence mortality

estimates. The correction of field data is important because larger carcasses are

detected by researchers more easily than smaller ones, and because the longer

time elapsed between consecutive searches and the smaller the size of the

carcasses, the larger the effect of scavengers on corpse disappearance [25].

Ideally, surveys to evaluate carcass losses should be carried out in each study

area before undertaking further mortality studies [25], because detection rates can

differ among study areas (e.g., due to habitat biases, [30]). Therefore, we used our

own correction equations instead of others recently published (e.g., [32]).

Observers were previously trained in order to minimize potential biases due to

their different levels of expertise in carcass searches [25].

63

Table 2. Equations from [25] used in our study to correct numbers of dead birds found at the

power line, in order to account for removal by scavengers or missed observations during carcass

searches. Different equations are given for the four size categories specified in [25] (see Table 3 for

their weights). We first corrected the number of carcasses found in the field by their size-

dependent detectability (A). Second, we applied equation B for different carcass sizes where “days”

is the number of days elapsed from the last visit. Third, we obtained a correction for every size

category. Finally, we added C to A to obtain the mortality estimates for each category. The mortality

estimate for a given power line was the sum of mortality estimates for the four carcass sizes.

Equation

An (Detectability)

A1 : Large= (no. carcasses found+1)*100/71.7 A2 : Medium= (no. carcasses found+1)*100/55.8 A3 : Small= (no. carcasses found+1)*100/32.1 A4 : Very small= (no. carcasses found+1)*100/33.3

Bn (Periodicity and scavenging)

B1 : Large = 0.744+28.063*log10(days) B2 : Medium=-1.751+41.880*log10(days) B3 : Small=-6.623+58.111*log10(days) B4 : Very small=13.538+60.342*log10(days)

Cn (Correction) (An*Bn)/100

Mortality estimate n An + Cn

In addition to testing the effectiveness of line marking as a means to reduce

bird collision rate, we also evaluated two potential sources of variation in marking

efficiency: power line type and spiral size. Whereas all transmission lines were

equipped with large spirals (35 cm diameter and 1 m length, Figure 1a), either

large or small spirals (10 cm of diameter and 24 cm m long, Figure 1b) were

attached to distribution lines, with the same spiral size attached to all the spans of

a given power line. We compared (i) the differences in marking efficiency in

transmission vs. distribution lines when equipped with large spirals; and (ii) the

efficiency of large vs. small spirals to reduce bird mortality in distribution lines.

64

Figure 1. Spirals used in our experiments. Difference in size between large (a) and small (b) can be

appreciated.

Unfortunately, we have no data on flight frequencies to estimate collision

rates associated with our different designs, but in the study of marking

effectiveness alone we used the corresponding controls to evaluate potential

changes in bird mortality associated with changes in bird population densities.

Furthermore, power lines of different categories were surveyed in the same study

area, minimizing the effect of potential local differences in bird densities.

Statistical analyses

As a basic first analytical approach we tested whether there was a trend in the

number of bird carcasses found after marking the line compared to before

marking. This was done considering each power line as a sample unit, and

comparing the number of decreases and increases in casualties recorded after

marking (in the case of experimental lines), or in the second survey year compared

to the first year (in the case of control lines). These comparisons were performed

using the two-tailed sign test for small samples [33]. The same test was carried out

using the total estimated number of dead birds, i.e. after correcting the number of

casualties recorded during the field surveys [25]. To confirm the observed trends,

we checked the differences in the accumulated numbers of estimated deaths

65

before-after marking (first-second year in the case of controls) and experimental

lines-control lines by means of a chi-squared test.

As a second approach we used a Generalized Linear Mixed Model (GLMM)

of various independent factors on the monthly estimated collision rate, after

applying corrections proposed by [25] to the number of carcasses found to account

for carcass losses due to removal by scavengers or to being overlooked by

observers. For this analysis we considered one month as a time lapse long enough

to allow the use of carcass search results in different months as statistically

independent. We performed three GLMMs with Poisson error distributions and log

link functions. The three analyses shared the same dependent variable, the

estimated number of dead birds per month, and standardizing per kilometre of

power line [30]. They also shared the random factor (power line). The models

were fitted by maximizing the log-likelihood using the Laplacian approximation in

R-Program 2.11.1 ([34]; lmer in lme4 package). The three analyses were the

following: (i) Marking effectiveness alone: We evaluated the effect of wire marking

on bird mortality with two fixed factors, ‘Marked vs. non-marked’, with two levels,

and ‘First survey year vs. second survey year’, also with two levels. This analysis

includes both lines marked in the second year, but not in the first, and control

lines. (ii) Power line type: We explored the effect of the power line type by

including a factor with two levels (transmission and distribution) in the sample of

power lines marked with large spirals. (iii) Spiral size: We studied the effect of

spiral size through a factor with two levels (large and small) in the sample of

distribution power lines.

In order to evaluate the importance of correcting for corpse losses, we

performed a sensitivity analysis with a second group of GLMM tests where the

dependent variable was the raw number of carcasses (i.e., those found in the field,

without correction per losses) per km per month. All other parameters remained

constant. This was only a methodological approach, as all the findings were based

on the above-mentioned estimated mortality.

Finally, to study the specificity of the patterns found, we re-analyzed our

data from a species-specific point of view. However, most of the species did not

allow analyzing them with a GLMM procedure because they were not well

66

represented in all the power lines along the study area. We thus proceeded with

Wilcoxon paired-sample tests for the three most common species: (i) doves (rock

and domestic doves and wood pigeons, all together), (ii) great bustards and (iii)

little bustards. We took into account the changes in mortality (first year vs. second

year) for the whole power line and separating experimental and control lines. We

made these species-specific calculations after correcting the number of casualties

recorded during the field surveys, i.e., with estimated mortality.

RESULTS

We found 521 carcasses of 45 bird species, 19 of conservation concern (Table 3).

Among experimental lines, most showed a decline in mortality after line marking

compared to before line marking (11 lines with a decrease, 4 with an increase; P=

0.10, two-tailed sign test). The overall decrease in the number of carcasses

recorded in the sample of 15 experimental lines was 88 birds (189 birds before

marking, 101 birds after marking, 47% reduction in observed casualties). In

control lines we did not observe a significant trend (10 lines with a decrease, 5

with an increase, 5 remained constant, P = 0.30, two-tailed sign test), with an

overall reduction of 20%.

67

Table 3. Species found dead under power lines in the present study and their size following [25]:

XS (<50g), S (50-150g), M (150-600g) and L (>600g). Figures are numbers of carcasses found

during the whole study period (2001-2010). Note that statistical analyses were made both with raw

data and after applying correction equations proposed by [25] to field data shown in this table. The

conservation status is based on [43] criteria: ‘SPEC 1’: European species of global conservation

concern; ‘SPEC 2’: Species having global populations concentrated in Europe and an unfavourable

conservation status in Europe; ‘SPEC 3’: species having global populations not concentrated in

Europe but an unfavourable conservation status in Europe; and, ‘Non-SPEC’: species having global

populations not concentrated in Europe and a favourable conservation status in Europe.

Species Size Carcasses found SPEC

Cattle Egret Bubulcus ibis L 9 Non-SPEC

White Stork Ciconia ciconia L 24 SPEC 2

Mallard Anas platyrhynchos L 4 Non-SPEC

Shoveler Duck A. clypeata L 1 Non-SPEC

Black Kite Milvus migrans L 2 SPEC 3

Cinereous Vulture Aegypius monachus L 2 SPEC 1

Marsh Harrier Circus aeruginosus L 1 Non-SPEC

Sparrowhawk Accipiter nisus M 1 Non-SPEC

Common Buzzard Buteo buteo L 1 Non-SPEC

Common Kestrel Falco tinnunculus M 6 SPEC 3

Red-legged Partridge Alectoris rufa M 10 SPEC 2

Common Quail Coturnix coturnix S 3 SPEC 3

Common Moorhen Gallinula chloropus M 2 Non-SPEC

Little Bustard Tetrax tetrax L 57 SPEC 1

Great Bustard Otis tarda L 73 SPEC 1

Stone Curlew Burhinus oedicnemus L 12 SPEC 3

Lapwing Vanellus vanellus M 19 Non-SPEC

Black-headed Gull Larus ridibundus L 2 Non-SPEC

Pin-tailed Sandgrouse Pterocles alchata M 6 SPEC 3

Rock/Domestic Dove Columba livia M 130 Non-SPEC

Wood Pigeon C. palumbus M 49 Non-SPEC

Common Swift Apus apus S 1 Non-SPEC

European Roller Coracias garrulus S 4 SPEC 2

Crested Lark Galerida cristata XS 1 SPEC 3

Skylark Alauda arvensis S 14 SPEC 3

Barn Swallow Hirundo rustica XS 1 SPEC 3

Meadow Pipit Anthus pratensis XS 7 Non-SPEC

Robin Erithacus rubecula XS 1 Non-SPEC

Northern Weather Oenanthe oenanthe XS 1 SPEC 3

Blackbird Turdus merula S 1 Non-SPEC

Reed Warbler Acrocephalus scirpaceus XS 1 Non-SPEC

Melodious Warbler Hippolais polyglotta XS 1 Non-SPEC

Subalpine Warbler Sylvia cantillans XS 3 Non-SPEC

68

Species Size Carcasses found SPEC

Orphean Warbler S. hortensis XS 1 SPEC 3

Blackcap S. atricapilla XS 2 Non-SPEC

Common Chiffchaff Phylloscopus collybita XS 4 Non-SPEC

Willow Warbler P. trochilus XS 3 Non-SPEC

Magpie Pica pica M 28 Non-SPEC

Jackdaw Corvus monedula M 1 Non-SPEC

European Starling Sturnus vulgaris S 1 SPEC 3

Spotless Starling S. unicolor S 8 Non-SPEC

House Sparrow Passer domesticus XS 3 SPEC 3

European Serin Serinus serinus XS 1 Non-SPEC

Linnet Carduelis cannabina XS 3 SPEC 2

Corn Bunting Emberiza calandra XS 7 Non-SPEC

Undetermined medium-sized bird M 3 ---

Undetermined passerine XS 6 ---

The 521 dead birds found represent 14,282 estimated bird collisions, an

average 8.2 collisions per month and km, after accounting for carcass removal by

scavengers and missed observations during surveys. Significantly more

experimental lines showed a decrease in the number of estimated casualties after

line marking compared to before line marking (12 lines with a decrease, 3 with an

increase; P= 0.04, two-tailed sign test). The overall difference in the sample of 15

lines was 316 birds (3,300 estimated birds before marking, 2,984 birds after

marking, 9.6% reduction in estimated mortality). The control sample did not show

significant before-after differences (10 lines with a decrease, 10 with an increase,

P= 1.0, two-tailed sign test; total estimated casualties: 4,067 before and 3,931 after

marking, 3.3% reduction). A chi-squared test with the former data (3,300, 2,984,

4,067 and 3,931) confirmed the difference between experimental and control

samples in the reduction of estimated casualties (2 = 3.90, P = 0.048).

69

In the GLMM considering all monthly surveys, the number of estimated

collisions per kilometre was significantly reduced in experimental power lines

after marking, while it remained similar in controls (Table 4i.a; Figure 2). This

model explained 96.4% of the deviance. The effectiveness of large spirals was

similar in transmission and distribution power lines (Table 4ii.a; Figure 3). The

model explained 99.6% of the deviance. Spirals of different sizes had similar

marking effectiveness when attached to distribution lines (Table 4iii.a; Figure 4),

with 98.8% of the deviance explained by the model. The comparisons with

uncorrected raw data (Table 4i.b, ii.b and iii.b) showed different statistical

differences (e.g., in ‘marked vs. non-marked’), highlighting the importance of

correcting field data.

Figure 2. Number of estimated carcasses per kilometre (mean ± SE) before (black) and after (grey

bars) marking in control (left) and experimentally marked (right) power lines. Sample sizes were

219 and 165 in each period for control and experimental power lines, respectively.

70

Figure 3. Number of estimated carcasses per kilometre (mean ± SE) before (black) and after (grey

bars) marking in transmission (left) and distribution (right) power lines. Sample sizes were 77 and

44 in each period for transmission and distribution power lines, respectively.

Table 4. Parameter estimates from the Generalized Linear Mixed Model for marking effectiveness

alone model (i), power line type model (ii) and spiral size model (iii). We show GLMM with (a) and

without (b) corrections for carcass losses due to researcher overlooking and removing by

scavengers. Estimate, standard error (SE), statistic value (z) and statistical significance (P) are

provided.

(i.a) Marking effectiveness alone (n=770) (with corrections)

Estimate SE z P

Intercept 2.34 0.09 27.31 <0.0001

Marked vs. non-marked -0.08 0.04 -2.13 0.03

First survey year vs. second survey year -0.04 0.03 1.57 0.12

(i.b) Marking effectiveness alone (n=770) (without corrections)

Estimate SE z P

Intercept -1.20 0.20 -6.35 <0.0001

Marked vs. non-marked -0.30 0.16 -1.90 0.06

First survey year vs. second survey year 0.47 0.14 3.46 <0.0001

(ii.a) Power line type (n=242) (with corrections)

Estimate SE z P

Intercept 2.10 0.11 18.49 <0.0001

Power line type 0.11 0.14 0.78 0.44

71

(ii.b) Power line type (n=242) (without corrections)

Estimate SE z P

Intercept -1.71 0.32 -5.42 <0.0001

Power line type 0.75 0.38 1.99 0.05

(iii.a) Spiral size (n=176) (with corrections)

Estimate SE z P

Intercept 2.10 0.08 25.12 <0.0001

Spiral size 0.10 0.12 0.88 0.38

(iii.b) Spiral size (n=176) (without corrections)

Estimate SE z P

Intercept -1.75 0.36 -4.92 <0.0001

Spiral size 0.65 0.49 1.32 0.19

Figure 4. Number of estimated carcasses per kilometre (mean ± SE) before (black) and after (grey

bars) marking in distribution power lines marked with large (left) and small (right) spirals. See

Figure 1 for more details. Sample sizes were 44 in all cases.

Regarding species-specific patterns, doves did not show significant

differences in the six treatments, regarding marking effectiveness alone (Wilcoxon

paired-sample test, marked vs. non-marked, Z = 0.87, P =0.39; first survey year vs.

second survey year, Z = 0.00, P =1.00), power line type (transmission lines, Z =

0.41, P =0.68; distribution lines, Z = 0.41, P =0.68) or spiral size (large spirals, Z = -

0.32, P =0.75; small spirals, Z = -0.50, P =0.62).

72

In contrast, great bustard mortality was reduced only after marking of

transmission lines (transmission lines, Z = 2.04, P =0.04; distribution lines, Z =

0.00, P =1.00) or only when marking with large spirals (large spirals, Z = 2.00, P

=0.046; small spirals, Z = -0.71, P =0.48), being not significant regarding marking

effectiveness alone (marked vs. non-marked, Z = 1.81, P =0.07; first survey year vs.

second survey year, Z = 0.00, P =1.00).

In the little bustard, wire marking reduced mortality (Z = 2.47, P =0.01),

whereas statistical differences were not found for controls (Z = 0.50, P =0.62) or

for power line type (transmission lines, Z = 1.79, P =0.07; distribution lines, Z =

1.15, P =0.25) or spiral size (large spirals, Z = 1.22, P =0.22; small spirals, Z = 0.00,

P =1.00).

DISCUSSION

Our results show a slight (overall, 9.6%, after correcting for carcass removal by

scavengers and missed observations), but significant reduction in bird mortality

after flight diverters were attached to power lines. Regardless of statistical

significance, a slight mortality reduction may be very biologically relevant in areas,

species or populations of high conservation concern. It is important to note that

overall mortality reduction values were not the same if calculated using raw

numbers of dead birds found, i.e. before correcting for carcass removal by

scavengers and missed observations. This is because correction factors differ

between species [25]. Thus, uncorrected mortality values would lead to incorrect

conclusions, and special care should be taken when dealing with certain birds of

conservation concern. Neither the type of line (transmission vs. distribution)

marked with large spirals, nor the size of spirals in distribution lines influenced the

magnitude of mortality reduction when we assessed overall mortality in all species

together. However, great bustard mortality showed reductions when lines were

marked with large spirals, and also considering only transmission lines.

The effectiveness of wire marking in reducing bird mortality through

collision has been recently reviewed by Barrientos et al. [19]. However, in that

study, different markers were combined since available sample sizes did not allow

73

inclusion of marker type as a factor in the analysis. Thus, despite spirals of

different sizes and colours being the most frequently employed bird flight

diverters, half of the studies included in Barrientos et al. [19] referred to other

device types (see Appendix in [19]). The present study suggests that the mortality

reduction found in that review was not due to the inclusion of other markers, and

that the most widely used spirals are effective. The present study also overcomes a

common problem detected in Barrientos et al. [19], namely that sample sizes are

generally small. Here we based our conclusions on a large sample including two-

year monthly surveys at 15 experimental and 20 control power lines, covering 72.5

km. Moreover, these lines were distributed over a relatively large geographical

area, encompassing most farmland areas used by steppe birds in our study region.

This overall low (9.6%) reduction could be greater in some places (e.g., migration

corridors, power lines close to resting sites, etc), or could represent a valuable

reduction for endangered species with high collision risk. Thus, a detailed

evaluation of mortality due to collision should be carried out before deciding

where to attach spirals as a bird protection measure in relatively large

conservation areas.

Some of the species found dead in our study are among those suggested in

previous studies to be the most likely to collide with power lines [2,8], namely

those with low maneuverability such as bustards, storks or waterfowl. These

species usually fly higher than, for instance, many passerines, and thus most of

their collisions are expected to be with transmission lines. Indeed, if we consider

the data from the first year only, i.e. before attaching spirals, transmission lines in

our study accumulated 71% (n=42) of all great bustards found dead in all lines,

50% (n=50) of all little bustards Tetrax tetrax, 83% (n=12) of all white storks

Ciconia ciconia and 100% (n=3) of all ducks Anas spp., despite the fact that

transmission lines represented only 36% of the total length of power lines

surveyed. In their study with whooping cranes, Morkill & Anderson [22] found that

birds reacted more than expected to transmission lines and less to distribution

lines. However, we did not find a significant difference in mortality reduction in

marked transmission lines compared to marked distribution lines when we

considered all species together. When looking at species-specific patterns, only the

great bustard showed a slight mortality reduction in marked transmission lines.

74

Although some studies found that species suffering high collision mortality may

show a tendency to avoid areas with transmission lines (e.g. little bustard, [35]),

collision with transmission lines is still one of the most important sources of

mortality in these species [35, 36]. Thus, as suggested in Barrientos et al. [19], it is

possible that at least some of these particularly sensitive species do not properly

respond to conventional marking methods (see below).

Although one would expect that large flight diverters are more effective

than small diverters in increasing the visibility of marked wires, other authors that

have used spirals of different sizes [23,24] did not statistically test for differences

among them. Our study explores this possibility for the first time. Considering all

species together, our results suggest that the decrease in collision rate is

independent of spiral size, and thus it seems reasonable to conclude that the main

advantage of marking is already achieved with small spirals, with larger spirals

being unnecessary. This could imply interesting applied findings. For example,

small diverters do not apply excessive weight to the wire. Large devices can

constitute a problem for this reason especially in high winds, contributing to the

downing of power lines, especially if devices are frozen [14,22]. However, a

flagship species like the great bustard showed mortality reduction with larger

spirals, suggesting that, at least for this species, large spirals work better.

Despite our study being, to our knowledge, the largest published field

experiment, and ca. 310,000 € were spent to mark 33.7 kilometres of power lines

in our study area, few conclusions can be drawn beyond the general effectiveness

of bird flight diverters in reducing collision mortality. We found differences in

effectiveness when we compared markers in transmission versus distribution

lines, or when we compared spirals of different sizes in distribution lines only with

one species (although we could carry out species-specific analyses only with three

species). However, it is worth noting that even after marking, bird collisions in our

study area were still high, especially for some endangered species usually showing

high collision risks (e.g. great and little bustards). Several non-mutually exclusive

explanations could account for this. First, it is possible that the generally low

probability of collision (0.21-0.05 birds per 1,000 crossings; [19]) makes it very

difficult to find differences even with well-designed experiments. If this is the case,

75

huge experimental designs would be necessary to find larger differences and

extract stronger conclusions. Second, it has been argued that bad weather or light

conditions can increase bird collisions, especially if birds have problems with flight

control [14,37]. For most birds, sustained slow flight is costly or aerodynamically

impossible [38, 39], and hence reducing speed is an unlikely mechanism to

increase safety under bad weather or light conditions. Third, collisions frequently

occur even under low wind and good visibility conditions [40]. Recent studies

[9,10] suggest that some species, which undergo high collision rates (e.g. bustards

and storks) have narrow fields of view in the frontal plane, hindering their ability

to see the way ahead. Fourth, Martin [10] suggests that birds flying in open

airspace above vegetation could relax –by means of either behavioural or

evolutionary adaptations- the monitoring of this airspace since it is a highly

predictable environment, usually clear of hazards. In other words, birds of some

species could simply not look ahead during flight. Indeed, frontal vision in birds is

not a high-resolution vision [10]. Instead, the best resolution occurs in the lateral

vision, which most birds employ to detect conspecifics (very important in social

species like bustards or storks) and predators, or in identify foraging

opportunities. All of these may be more important for a bird than simply looking

ahead during flight into open airspace [10]. Fifth, anecdotal events can have

potentially important effects on collisions. For instance, Sastre et al. [41] suggest

that human-related disturbances causing flight response can increase the

probability of collision of great bustards with power lines. Sixth, regarding the

effectiveness evaluation of different devices, it is also plausible that misguided

approaches have been used to date. For instance, whereas bird flight diverters are

usually coloured with a single colour bright to the human eye [19], a recent review

[10] recommends the use of black-and-white diverters, which reflect highly or

absorb strongly across the full spectrum of ambient light. Thus, it is possible that

the few valuable studies carried out to date that compared the effectiveness of

different colours for a certain bird flight diverter [42] actually compared colours

too close in the spectrum to identify differences in their effectiveness. Since it is

recognized that the colour vision of birds extends into the ultraviolet range, thus

broadening, compared with humans, the range of stimuli to which the avian eye

can respond [10], the use of ultraviolet-devices should be investigated.

76

In summary of the above-mentioned explanations, and given that is seems

clear that no single type of marker will be equally effective for all bird species, we

acknowledge that the importance of type and size of bird flight diverters is not yet

clear and should be confirmed in future studies. Our study does not pretend to be

comprehensive in this respect, and regarding the different susceptibilities of

different bird species or groups to collision [see 2,8], and particularly the mortality

reductions obtained for specific models of flight diverters should be further

investigated. In this sense, we encourage researchers to explore the effectiveness

of non-visual diverters. Finally, we highly recommend the identification of

mortality hot-spots based on the number of individuals killed and the vulnerability

of the species involved [e.g. 44]. Taking into account the economic cost of marking,

it is likely more useful to attach flight diverters to these hot-spots rather than to do

it to whole sections of power line.

ACKNOWLEDGEMENTS

We thank A. García Fernández and M. Carrasco for their assistance during the field

work. We also thank J. Camaño and J. Velasco of HENARSA, and the electric

companies Iberdrola, Unión Fenosa and Red Eléctrica de España for their

cooperation. S. Young reviewed the English.

REFERENCES

1. APLIC (Avian Power Line Interaction Committee) (2006) Suggested practices for

avian protection on power lines: state of the art in 2006. Washington:

Edison Electric Institute and Sacramento: California Energy Commission.

227 p.

2. Bevanger K (1998) Biological and conservation aspects of bird mortality caused

by electric power lines. Biol Conserv 86:67-76.

77

3. Jenkins AR, Smallie JJ, Diamond M (2010) Avian collisions with power lines: a

global review of causes and mitigation with a South African perspective.

Bird Conserv Int 20: 263-278.

4. Manville AM II (1999) The ABC’s of avoiding bird collisions at communication

towers: the next steps. In: Avian interactions with utility and

communication structures. Workshop proceedings. Palo Alto: Electric

Power Research Institute. pp 85-104.

5. Manville AM II (2009) Towers, turbines, power lines, and buildings: steps being

taken by the U.S. Fish and Wildlife Service to avoid or minimize take of

migratory birds at these structures. In Ralph CJ, Rich, editors. Tundra to

tropics: connecting birds, habitats, and people. Proceedings 4th

international Partners in Flight conference. McAllen: Partners in Flight. pp

262-272.

6. Koops FBJ (1994) Collision victims of high-tension lines in The Netherlands and

effects of marking. In: Red Eléctrica Española. First technical sessions on

power lines and the environment. Madrid: REE. pp 51-57.

7. Hunting K (2002) A roadmap for PIER research on avian collisions with power

lines in California. Technical report P500–02-071F. Sacramento: California

Energy Commission. 69p.

8. Janss GFE (2000) Avian mortality from power lines: a morphologic approach of a

species-specific mortality. Biol Conserv 95: 353–359.

9. Martin GR, Shaw JM (2010) Bird collisions with power lines: failing to see the

way ahead? Biol Conserv 143: 2695-2702.

10. Martin GR (2011) Understanding bird collisions with man-made objects: a

sensory ecology approach. Ibis 153: 239-254.

11. Bevanger K (1995) Estimates and population consequences of tetraonid

mortality caused by collisions with high tension power lines in Norway. J

Appl Ecol 32: 745-753.

78

12. Schaub M, Pradel R (2004) Assessing the relative importance of different

sources of mortality from recoveries of marked animals. Ecology 85: 930-

938.

13. Shaw JM (2009) The end of the line for South Africa’s national bird? Modelling

power line collision risk for the Blue Crane. MS thesis. Cape Town:

University of Cape Town.

14. Beaulaurier DL (1981) Mitigation of bird collisions with transmission lines.

Portland: Bonneville Power Administration. 105 p.

15. Bevanger K, Brøseth H (2001) Bird collisions with power lines: an experiment

with ptarmigan (Lagopus spp.). Biol Conserv 99: 341–346.

16. APLIC (Avian Power Line Interaction Committee) (1994) Mitigating bird

collisions with power lines: the state of the art in 1994. Washington: Edison

Electric Institute. 78 p.

17. Hebert E, Reese E, editors. (1995) Avian collision and electrocution: an

annotated bibliography. Publication P700–95-001. Sacramento: California

Energy Commission. 115 p.

18. Alonso JC, Alonso JA, Muñoz-Pulido R (1994) Mitigation of bird collisions with

transmission lines through groundwire marking. Biol Conserv 67: 129-134.

19. Barrientos R, Alonso JC, Ponce C, Palacín C (2011) Meta-analysis of the

effectiveness of marked wire in reducing avian collisions with power lines.

Conserv Biol 25: 893-903.

20. Kruger R (2001) A risk based approach to the cost of implementing raptor

mitigation measures on Eskom distribution networks in South Africa. In:

Carlton R, project manager. Avian interactions with utility and

communication structures. Workshop proceedings. Palo Alto: Electric

Power Research Institute. pp 229-246.

21. Alonso JC, Martín CA, Palacín C, coordinators. (2005) Proyecto de medidas

preventivas, correctoras y compensatorias de la afección de la M-50 y de la

Autopista de Peaje R-2 a la población de avutardas y otras aves esteparias

79

de la IBA Talamanca-Camarma, y al LIC Cuenca de los ríos Jarama y

Henares. Madrid: Museo Nacional de Ciencias Naturales (CSIC). 37 p.

22. Morkill AE, Anderson SH (1991) Effectiveness of marking powerlines to reduce

sandhill crane collisions. Wildlife Society Bulletin 19: 442-449.

23. Koops FBJ, de Jong J (1982) Vermindering van draadslachtoffers door

markering van hoogspanningsleiden in de omgeving van Heerenveen.

Elektrotechniek 60: 641–646.

24. Anderson MD (2002) The effectiveness of two different marking devices to

reduce large terrestrial bird collisions with overhead electricity cables in

the eastern Karoo, South Africa. Report 1. Johannesburg: Eskom. 25p

25. Ponce C, Alonso JC, Argandona G, Fernández AG, Carrasco M (2010) Carcass

removal by scavengers and search accuracy affect bird mortality estimates

at power lines. Anim Conserv 13: 603-612.

26. Ponce C, Bravo C, García de León D, Magaña M, Alonso JC (2011) Effects of

organic farming on plant and arthropod communities: A case study in

Mediterranean dryland cereal. Agric Ecosyst Environ 141: 193–201.

27. Alonso JC, Martín CA, Palacín C, Magaña M, Martín B (2003) Distribution, size

and recent trends of the Great Bustard (Otis tarda) population in Madrid

region, Spain. Ardeola 50: 21–29.

28. García de la Morena EL, Bota G, Ponjoan A, Morales MB (2006) El sisón común

en España. I Censo Nacional (2005). Madrid: SEO/BirdLife. 159p.

29. Martín CA, Palacín C, Martín B, Ponce C, Sastre P, Bravo C (2011) Abundancia y

distribución de la ganga ortega (Pterocles orientalis) y la ganga ibérica

(Pterocles alchata) en la Comunidad de Madrid. Anuario Ornitológico de

Madrid 2009-2010. In press.

30. Bevanger K (1999) Estimating bird mortality caused by collision and

electrocution with power lines; a review of methodology. In: Ferrer M, Janss

GFE, editors. Birds and power lines. Collision, electrocution and breeding.

Madrid: Quercus. pp. 29-56.

80

31. Virgós E, Travaini A (2005) Relationship between small-game hunting and

carnivore diversity in central Spain. Biodiv Conserv 14: 3475-3486.

32. Huso MMP (2011) An estimator of wildlife fatality from observed carcasses.

Environmetrics 22: 318-329.

33. Siegel S, Castellan NJ (1988) Nonparametric statistics for the behavioural

sciences, 2nd edn. McGraw-Hill, New York

34. R Development Core Team (2009) R: A language and environment for

statistical computing. R Foundation for Statistical Computing, Vienna.

Available at: http://www.R-project.org. Last accessed: 17/01/2012.

35. Silva JP, Santos M, Queirós L, Leitão D, Moreira F, et al. (2010) Estimating the

influence of overhead transmission power lines and landscape context on

the density of little bustard Tetrax tetrax breeding populations. Ecol Model

221: 1954-1963.

36. Martín B (2008) Dinámica poblacional y viabilidad de la avutarda común en la

comunidad de Madrid. PhD Thesis. Universidad Complutense de Madrid,

Madrid.

37. Savereno AJ, Savereno LA, Boettcher R, Haig SM (1996) Avian behavior and

mortality at power lines in coastal South Carolina. Wild Soc Bull 24: 636-

648.

38. Norberg UM (1990) Vertebrate Flight: Mechanics, Physiology, Morphology,

Ecology and Evolution. Zoophysiology Series, Vol. 27. Berlin: Springer

Verlag. 291p.

39. Biewener AA (2003) Animal Locomotion. Oxford: Oxford University Press.

281p.

40. Drewitt AL, Langston RHW (2008) Collision effects of wind-power generators

and other obstacles on birds. Annals of the New York Academy of Sciences

1134: 233-266.

81

41. Sastre P, Ponce C, Palacín C, Martín CA, Alonso JC (2009) Disturbances to great

bustards (Otis tarda) in central Spain: human activities, bird responses and

management implications. Eur J Wildl Res 55: 425-432.

42. Calabuig CP, Ferrer M (2009) Análisis de la eficacia y la vida útil de la

señalización anticolisión “salvapájaros” en líneas de transporte de energía

eléctrica. Seville: REE, SAU, and CSIC. 41p.

43. BirdLife International (2004) Birds in Europe: population estimates, trends

and conservation status. Cambridge: BirdLife International. 374p.

44. Quinn M, Alexander S, Heck N, Chernoff G (2011) Identification of bird collision

hotspots along transmission power lines in Alberta: an expert-based

geographic information system (GIS) approach. J Environ Inform 18: 12-21.

82

83

CAPÍTULO 3

84

Este capítulo reproduce íntegramente el siguiente artículo:

Ponce, C., Bravo, C., García de León, D., Magaña, M. & Alonso, J.C. 2011. Effects of

organic farming on plant and arthropod communities: a case study in

Mediterranean dryland cereal. Agriculture, Ecosystems & Environment 141: 193–

201.

85

CAPÍTULO 3

Effects of organic farming on plant and

arthropod communities: a case study in

Mediterranean dryland cereal

Carlos Poncea, Carolina Bravoa, David García de Leónb, Marina

Magañaa, Juan Carlos Alonsoa

a Dep. Ecología Evolutiva, Museo Nacional de Ciencias Naturales, CSIC, José

Gutiérrez Abascal 2, E-28006 Madrid, Spain

b Dep. Protección de Cultivos, Instituto de Agricultura Sostenible, CSIC, Finca

Alameda del Obispo, E- 14080 Córdoba, Spain

86

ABSTRACT

Organic farming is considered an important way to preserve biodiversity in agricultural

landscapes. However, more work is still necessary to enable a full appraisal of the

potential benefits of this way of farming, since studies differ in the evaluation of its

effectiveness. Studies are particularly scarce in the Mediterranean region, where different

climatic and ecological conditions prevent simple extrapolations from work carried out at

northern latitudes. In the present study, an analysis of weed and arthropod communities

was conducted in 28 pairs of organic and conventional fields in a dry cereal farmland in

central Spain. Plants were identified to the species level, and arthropods to the family

level. Pitfalls and sweep nets were used to sample respectively, ground-dwelling and

plant-visiting arthropods. Abundance (total numbers of individuals), richness (total

numbers of plant species or arthropod families), diversity (Shannon-Wiener index) and

biomass (milligrams per pitfall/sweep-net) were calculated for each field and compared

between organic and conventional fields using Generalized Linear Mixed Models (GLMMs).

To explore the effect of predictor variables on weed richness and arthropod biomass,

GLMMs were used. Organic fields showed higher abundance of weeds and arthropods

(respectively, 3.01 and 1.43 times), higher weed richness and diversity (respectively, 2.76

and 2.33 times), and a 24% reduction in cereal plants. Arthropod diversity was lower in

organic fields due to the presence of three dominant groups: Collembola, Chloropidae

(Diptera), and Aphididae (Hemiptera). Weed richness increased as cereal cover decreased

in organic fields. Total arthropod biomass was slightly higher in organic fields, and was

affected by weed abundance and diversity. The differences between organic and

conventional fields found in this study were higher than those reported for northern

latitudes. This could be explained by the richer weed flora in the Mediterranean region,

and a higher weed seed availability favoured by the two-year rotation system typical of

Iberian dry cereal farmland. We conclude that organic farming may contribute to preserve

biodiversity in dryland cereal agroecosystems in the Mediterranean region.

Keywords

Diversity, richness, abundance, weed and arthropod, agri-environment scheme,

farmland.

87

INTRODUCTION

A wealth of evidence points to agricultural intensification as the main cause of

biodiversity loss in farmland ecosystems (Donald et al., 2006; Foley et al., 2005;

Millennium Ecosystem Assessment, 2005; Wilson et al., 2009, 2010). This negative

impact of modern agriculture on many plant and animal taxa will probably raise in

the future, due to increasing demands in agricultural production. This is at present

an issue of major concern worldwide (Clough et al., 2007a; Fuller et al., 2005; Hole

et al., 2005), and there is a growing consensus that further increases in agricultural

production must avoid further adverse environmental impacts (Firbank, 2009;

Royal Society, 2009). One of the ways to reverse this negative trend would be to

use organic farming methods (Geiger et al., 2010). Agri-environment schemes

including organic farming and other environmentally friendly practices are today

considered the most important instruments to counteract the negative effects of

modern agriculture (EEA, 2004). However, published studies differ in their

evaluation of the effectiveness of these measures, which makes it difficult to assess

their benefits (Bengtsson et al., 2005; Frampton and Dorne, 2007; Kleijn et al.,

2006).

In a comprehensive review of comparative studies of organic and

conventional farming systems, Hole et al. (2005) found inconsistencies between

and within studies which suggested that the benefits to biodiversity of organic

farming may vary according to factors such as location, climate, crop-type and

species. They concluded that further studies are still needed in order to

understand the impacts of organic farming, before a full appraisal of its potential

role in biodiversity conservation in agroecosystems can be made. For example,

many recent studies have attempted to evaluate the effectiveness of organic

farming using birds, plants or invertebrates as study subjects (Beecher et al., 2002;

Bengtsson et al., 2005; Diekötter et al., 2010; Chamberlain et al., 2010; Clough et

al., 2005; Clough et al., 2007a; Fuller et al., 2005; Gabriel et al., 2006; Gabriel et al.,

2010; Gibson et al., 2007; Piha et al., 2007; Roschewitz et al., 2005; Schmidt et al.,

2005; Weibull et al., 2003). However, most of these studies have been carried out

at mid- or high latitudes of the northern hemisphere, and very few in the

Mediterranean Region, where climatic conditions are quite different (e.g., lower

88

rainfall, higher temperatures, lower soil organic content, and considerable

variation in the amount of water available for different springs; Costa et al., 2004;

INE, 2009; Walter, 1994), making it difficult to extrapolate the conclusions from

northern latitudes (Hole et al., 2005).

Two recent studies address the effect of organic farming on

arthropods in the Mediterranean Region, but not in dryland cereal fields (Cotes et

al., 2010; Hadjicharalampous et al., 2002). In this Region, only three studies used

vascular plants as study subjects. In José-María et al.´s (2010) study, management

was the main factor explaining differences among field centres, while Romero et al.

(2008) found that organic farming increased weed cover, and species richness and

diversity. Another study carried out in four organically managed fields (Caballero-

López et al., 2010) showed that plants are highly dependent on farming system,

and the arthropod community is conditioned by those plants, which led the

authors to conclude that interactions are also important in order to assess the

importance of management in cereal crops. Finally, in their recent review, Hole et

al. (2005) stated the need of further studies particularly in the Mediterranean

region.

In the present study we evaluated the effects of organic farming on

biodiversity in a dry cereal farmland in central Spain. The aim was to determine

whether there were any differences in the weed and arthropod communities

between fields that had been farmed without using synthetic fertilizers and

pesticides (organic system), and fields where these chemicals were used

(conventional system). Therefore, unlike most previous studies that concentrated

on single plant or invertebrate groups, we quantified the effect of the agro-

chemical treatment on the abundance (total numbers of individuals), richness

(total numbers of species or families), and diversity (Shannon-Wiener index) of all

identifiable vascular plants and arthropods found, as well as on the cover of grown

cereal and weeds, and on arthropod biomass. Besides, we characterized the factors

affecting both weed richness and arthropod biomass, since these are some of the

most studied variables in organic farming studies.

89

METHODS

Study area, field selection and farming practices

The study was conducted in 2008 in a Special Protection Area for birds (SPA 139,

‘Estepas Cerealistas de los ríos Jarama y Henares’) about 25 km north of Madrid

(40º42'N, 3º29'E; 682 m.a.s.l.), in central Spain. The terrain is flat to slightly

undulated, and it is primarily dedicated to dryland cereal cultivation (wheat

Triticum aestivum (L.), barley Hordeum vulgare (L.), and smaller amounts of

common oat Avena sativa (L.), together more than 95% of the surface), with minor

fields of legumes (Vicia spp. and Medicago sativa (L.)), olive groves Olea europaea

(L.) and grapevines Vitis vinifera (L.). The brown and acid soil present in the study

area and the weather conditions favor a natural vegetation composed by evergreen

oak trees (Quercus ilex (L.); and their degraded states –Retama sp. and Thymus sp.

scrubland-), which instead of forming dense woods have been cleared up to open-

wooded area called ‘dehesas’ used for wood extraction and livestock grazing.

Scattered groups of white poplars (Populus alba (L.)) are also found in the SPA,

although as in the case of oaks, always more than 1 km away from our sampling

fields, and thus probably having no influence on them. Most cereal is grown in a

traditional two-year rotation system, and harvested during late June-early July.

The climate is Mediterranean, with an annual precipitation (mean ± S.D.) of 442.5 ±

125.5 mm and a mean annual temperature of 14.4 ºC (maximum and minimum

temperatures, respectively, 42.2 ºC and -14.8 ºC). During the study year, the mean

annual precipitation was 484.9 mm and the mean monthly temperature, 14.3 ºC

(maximum and minimum temperatures, respectively, 39.3 ºC and -6 ºC). The mean

temperature during May is 15.6 ± 1.6 ºC, and the mean rainfall, 55.1 ± 41.2 mm. In

May 2008 these values were, respectively, 15.5 ºC and 64.7 mm, so we can

consider our study year as normal. The study area is a SPA for birds because it

holds significant populations of globally threatened steppe birds. Therefore, an

agri-environmental scheme is running in this area since 2001, as part of the

compensatory measures for the construction of a highway crossing its southern

margin. Organic farming was one of the conservation actions implemented in a

sector of the SPA.

90

Twenty-eight pairs of fields were randomly selected, where one field of

each pair was cultivated without synthetic fertilizers and pesticides (organic

system), and the other field with such products (conventional system) (see e.g.

Clough et al., 2007b; Pfiffner and Niggli, 1996; Shah et al., 2003). All sampled cereal

fields (always dedicated to cereal cultivation) were preceded by a fallow year

before the study was carried out, so the initial conditions were the same for all of

them and the only difference was that one field of the pair was cultivated

organically during the year when our study was conducted. Fields of the same pair

were separated by <100 m and shared the major physiographic characteristics

(slope, orientation, approximate size, soil type) and farm history. The mean field

size was 1.9 ± 0.9 ha, similar to that of a previous study in northern Spain (José-

María et al., 2010). Since the maximum distance between fields in our sample was

11 km, we considered that the environmental conditions were the same for all

fields.

Farmers were asked to fill out a questionnaire to characterize their usual

farming practices, which are compared to those allowed in organic fields (Table 1).

Both organic and conventional fields were sown (wheat or barley) between the

second week of October and the first week of November 2007, after initial

ploughing for soil preparation. Conventional fields were later treated with

chemical fertilizers (Table 1) and broad-leaf herbicides, while organic fields did

not receive such treatments. The density of seeds (wheat or barley) was the similar

in both, organic and conventional fields (T = 1.80, P = 0.12, Table 1).

91

Table 1. Main characteristics of the farming system used in the 28 pairs of fields.

Organic fields Conventional fields

Sowing density (wheat or barley)

188 ± 16 kg ha-1 197 ± 19 kg ha-1

Fertilization No NPK: 350 ± 72 kg ha-1, October

CAN (27%): 168 ± 26 kg ha-1, February

Weed control Weed ploughing Weed ploughing

Clorsulfuron (7%): 2-2.5 g ha-1. April, May and July

Clortoluron (50%): 3-4 l ha-1

Gardel: 0.2 l ha-1

Foramsulfuron: 10 g ha-1. April, May and July

Primafuron: 20 g ha-1

Seed origin Organic Industrially selected and chemically treated

Ploughing (mouldboard plus weed ploughing)

1-2 times/year 2-4 times/year

Plant and arthropod sampling

Plant sampling was carried out during the third week of May 2008. A 25 x 25 cm

metal square was thrown randomly 20 times in each field, avoiding the edges and

their proximities. Each plant was identified to the species level, the number of

individual plants of each species was counted and the corresponding cover for

each species estimated as a percent of the square surface. In the case of cereal, the

total number of plants was used as an indirect measure of cereal production, since

no information about crop could be obtained. To check if plant sampling effort was

sufficient, species accumulation curves were generated using the program

EstimateS version 8.2 (Colwell, 2009) and fitted by Clench equation (Jiménez-

Valverde et al., 2003; Moreno and Halffter, 2001). The Clench equation was defined

as Sn=AxN/(1+BxN), where Sn is the number of species observed in each given

sample level, A is the increase rate of new species at the beginning of sampling and

B is the parameter related to shape of the curve. The asymptote of curve -total

number of species predicted- is calculated as A/B.

We used two different methods to sample arthropods, pitfall traps and

sweep nets. Pitfall traps are the most appropriate method to capture terrestrial

92

and soil arthropods (e.g., Clere and Bretagnolle, 2001; Hadjicharalampous et al.,

2002; Schmidt et al., 2006), while sweep nets are commonly used for taxa such as

Heteroptera that are living well above the ground on the plant canopy, or those

that spend much time flying (Frampton and Dorne, 2007). The combination of both

methods provides the best possible information about the arthropod fauna

(Fauvel, 1999). Within each field, three pitfall traps were placed during the third

week of May close to the field center at 10-m intervals. Each trap consisted of a

plastic cup (9 cm internal diameter, 14 cm length) sunk into the soil with the aid of

a metal cylinder, and filled with 250 ml of 70% ethanol as a preservative solution

(Shah et al., 2003). Traps were protected from rainfall and excessive evaporation

by plastic dishes suspended on thin sticks at 10 cm over the soil surface.

Collections of arthropods were made for 7 days (± 2 hours). Collected arthropods

were stored in 70% ethanol after the sampling period. Five days after collecting

the pitfalls, we conducted three sweep-netting transects on each field. Fields of the

same pair were sampled one after each other, between 6:00 h and 10:00 h GMT,

avoiding inappropriate weather conditions such as wind and temperature below

18°C or above 25°C, when arthropods might be inactive (Weibull and Östman,

2003). Each transect consisted of ten movements of the sweep net, from right to

left side and vice versa and approximately 2 m wide. Before starting these

samplings, all observers spent one day standardizing these sweep net movements

to prevent sampling biases due to differences in width, depth and speed, and the

same observer always sampled both fields of a pair. The arthropods captured were

fixed in 70% ethanol.

Laboratory procedures and statistical analyses

Plants were identified to the species level and arthropod to the family level, which

is useful for all indexes used in this study (abundance, richness and Shannon-

Wiener diversity index; Biaggini et al., 2007; Frampton and Dorne, 2007), as well

as for biomass calculations (Hódar, 1996). To estimate arthropod biomass we first

measured with a digital caliper (0.01 mm precision) the maximum body length of

all adult arthropods captured excluding appendices (wings, antennae, ovipositors

or legs), and calculated the average body size for each taxonomic group. To

estimate the mean biomass of each group, we used the equations given by Hódar

93

(1996), which relate weight to body length in several arthropod groups of the

Mediterranean region (general equation: Y = ab1(x)b2, where Y is the biomass, x the

length, and a,b1 and b2, specific coefficients for each taxonomic group).

Consequently we calculated the biomass of each group in each sampled field (see

also Clere and Bretagnolle, 2001; Jiguet et al., 2000).

We compared each index (richness, abundance, diversity and biomass) by

means of Generalized Linear Mixed Models (GLMMs) with the field pair as random

factor (to control for spatial non-independence in the data; Littel et al., 2006) using

the lme4 package (Bates and Maechler, 2010) of R-Program 2.11.1 (R Development

Core Team, 2010).

The differences in frequency distribution of the most abundant weed

species between organic and conventional management were analyzed using Chi-

squared test. Richness, abundance and diversity (plus biomass for arthropods)

were calculated independently for plants and for arthropods. Later, these indices

were calculated separately for cereal plants, weeds, and all plants (Kleijn et al.,

2006; Lundkvist et al., 2008; Sunderland and Samu, 2000). Finally, we calculated

these indices again and repeated the GLMMs for the most abundant arthropod

orders. To check for dominant groups, we used the index proposed by Berger and

Parker (1970). This index accounts for the dominance of the most abundant

groups (the higher the value, the more dominant group), considering all species in

the assemblage (Caruso et al., 2007). We repeated the diversity calculations

excluding dominant arthropod groups.

Since arthropod biomass is expected to be related to vegetation variables

(e.g., Clough et al., 2007b), we performed simple correlations analysis to discard, if

necessary, some highly correlated variables. Next, we performed Generalized

Linear Mixed Models (GLMMs), using field pair as random parameter, with Poisson

error distribution and log link function. Biomass was the dependent variable and

we included management type (organic or conventional), weed abundance, weed

richness, weed diversity and weed cover as plausible independent variables. We

performed another GLMM (with field pair as random effect) where weed richness

was the dependent variable, and cereal cover and management, plus their

interaction, the explanatory variables.

94

To determine the best predictive models, Akaike’s information criterion

(ΔAICc <2) was used. We used AICc because the ratio between the number of

observations and estimator variables was under 40 (Barrientos and Bolonio 2009;

Burnham and Anderson 2002). To look for differences among models with ΔAICc <

2, an ANOVA test was performed. The models were fitted by maximizing the log-

likelihood using the Laplacian approximation because this is the most suitable for

small sample sizes (Moya-Laraño and Wise 2007).

RESULTS

Plants

A total of 4940 plants belonging to 51 weed species were recorded, (Appendix A).

The frequency distribution of these species differed between organic and

conventional fields (2 = 8467.2, P < 0.001; Table 2). Only four weeds were found

in organic fields in lower numbers than in conventional fields (Lolium rigidum

(Gaudin), Avena sterilis (L.), Polygonum aviculare (L.), and Filago lutescens (Jord.);

Table 2). According to the Clench equation, we sampled 80% and 89.8% of the

total number of predicted species, respectively in organic and conventional fields.

The most abundant family was Gramineae, with 71.1% of total weeds (respectively,

60.9% and 88.1% in organic and conventional fields). Next were Compositae, with

10.4% (respectively, 14.4% and 3.8% in organic and conventional fields) and

Leguminosae, with 4.3% (respectively, 7.3% and 0.2% in organic and conventional

fields). Of 51 weed species identified, 48 were found in organic fields and 28 in

conventional fields.

95

Table 2. Most common weed species ordered by the frequency with which they were recorded in

organically managed and conventional fields.

Species Organic fields Conventional fields

Lolium rigidum (Gaudin) 50.3 80.2

Galium tricornutum (Dandy) 7.1 1.8

Bromus diandrus (Roth) 6.6 4.2

Anacyclus clavatus (Desf.) 4.2 0.7

Conyza canadensis (L.) 4.0 0.6

Raphanus raphanistrum (L.) 3.8 0.6

Avena sterilis (L.) 3.0 3.3

Vicia sativa (L.) 2.8 0.0

Polygonum aviculare (L.) 2.6 2.9

Filago pyramidata (L.) 1.7 0.1

Trifolium angustifolium (L.) 1.7 0.0

Filago lutescens (Jord.) 1.1 1.2

Vicia spp. 0.9 0.1

Picnomon acarna (L.) 0.9 0.8

Lactuca serriola (L.) 0.8 0.2

Ornithopus compressus (L.) 0.7 0.1

Linaria viscosa (L.) 0.7 0.0

Spergula arvensis (L.) 0.7 0.0

Euphorbia serrata (L.) 0.6 0.0

Vicia ervilia (L.) 0.6 0.0

Others 4.9 3.2

Values are percentages of each species found in both field types.

GLMMs showed that weed richness, weed diversity, weed abundance and

weed cover were significantly higher in organic than conventional fields (Table 3),

whereas cereal plants grew in higher numbers in conventional fields (Table 3).

Total plant abundance (cereal plus weeds) was higher in conventional fields, and

cereal cover and total cover did not differ between organic and conventional fields

(Table 3). Overall, there was a negative relationship between cereal cover and

weed richness, although this relationship was only significant for organically

managed fields (Fig. 1). The GLMM showed that weed richness was influenced by

the management type, cereal cover and their interaction (Appendix B, Table 4).

96

The first two models are equally valid (ANOVA test not significant), but the

first including the interaction and had a lower AICc.

Table 3. Differences between conventional and organic fields in abundance, cover, richness, and

diversity of plants.

Organic fields Conventional fields Z P LL

Abundance

Cereal 475.2 ± 242.6 623.6 ± 141.2 -23.6 <0.001 -953.2

Weeds 132.5 ± 153.8 43.9 ± 75.5 35.38 <0.001 -1214

Both 607.7 ± 241.9 667.5 ± 125.4 -7.69 <0.001 -785

Cover

Cereal 23.5 ± 14.2 35.7 ± 12.2 -0.39 0.261 -110.8

Weeds 9.9 ± 9.8 1.9 ± 2.8 12.59 <0.001 -73.6

Both 33.4 ± 14.1 37.6 ± 11.2 -0.89 0.143 -104.5

Richness Weeds 9.4 ± 4.0 3.4 ± 1.8 8.63 <0.001 -36.60

Diversity Weeds 1.4 ± 0.5 0.6 ± 0.5 3.05 0.002 -13.25

Abundance measured as individuals per 1.25 m2, or twenty 25x25cm sampling units, cover as %,

richness as species per 1.25 m2, and diversity through Shannon-Wiener diversity index. Mean

values ± SD, statistic (Z, GLMM-test), significance of the differences (P) and log-likelihood (LL) are

given.

Figure 1. Relationship between cereal cover (%) and weed richness (no. of species per 1.25 m2).

The correlation was significant for organic fields (open circles; r = -0.62, P < 0.001), but not for

conventional fields (black circles; r = -0.23, P = 0.23).

97

Table 4. Parameter estimates from the Generalized Linear Mixed Model with cereal cover (CC),

management (MA), and the interaction between cereal cover and management (CC*MA) as factors

affecting weed richness.

Parameter Estimate SE P

CC -0.009 0.052 0.452

MA 6.481 2.301 0.007

CC*MA -1.873 0.325 0.031

Arthropods

A total of 82822 individuals belonging to 150 arthropod families and 21 orders

were collected (Appendix C). Arthropods were more abundant in organic fields

(50488 individuals, Table 5).

Table 5. Differences between conventional and organic fields in abundance, richness, diversity, and

biomass of arthropods.

Organic fields Conventional fields Z P LL

Abundance 1798.4 ± 1052 1253.9 ± 508 54.66 <0.001 -3861

Richness 45 ± 4.9 42.3 ± 4.29 6.98 0.002 -26.98

Diversity 1.9 ± 0.1 2.2 ± 0.2 -2.57 0.014 -13.61

Biomassa 719 ± 153 640 ± 275 1.82 0.08 -16.2

a one pitfall plus one sweep net transect.

Abundance measured as individuals collected in 3 pitfalls/sweep-nets, richness as number of

families collected in 3 pitfalls/sweep-nets, diversity through Shannon-Wiener diversity index, and

biomass of arthropods as milligrams in 1 pitfall/sweep-net. Mean values ± SD, statistic (Z, GLMM-

test), significance of the differences (P) and log-likelihood (LL) are given.

Comparisons between organic and conventional fields showed no

significant differences for Araneae, Coleoptera or Hemiptera (P > 0.41 in all cases),

higher numbers of Acari, Collembola, Diptera, Hymenoptera, and Orthoptera in

organic fields (P < 0.001), and higher numbers of Thysanoptera in conventional

fields (P < 0.001) (Fig. 2). Dominance analyses showed that Collembola,

Chloropidae (Diptera), and Aphididae (Hemiptera) were dominant groups (Berger-

98

Parker index = 0.68 for these three groups together; respectively 0.76 and 0.58 in

organic and conventional fields).

Figure 2. Numbers of individuals of the main arthropod orders sampled per field (three pitfalls

plus three sweep nets transects), in organic (open circles) and conventional fields (black circles).

Abundance measured as ln (number of individuals + 1). Means and SD values are given.

The result of the GLMMs showed that richness was higher in organic than in

conventional fields (Table 5). After excluding dominant groups, richness was still

higher in organic (40.3 ± 1.2) than in conventional fields (37.8 ± 1.1) (P = 0.02). No

differences were found in family richness for Araneae and Coleoptera (respectively,

P = 0.56 and P = 0.11, see list of families in Appendix C). For Hemiptera, richness

was higher in organic fields (P = 0.01). As for Diptera, richness did not differ

between organic and conventional fields (P = 0.95).

Diversity values were lower in organic than in conventional fields (Table 5).

Within the most abundant orders, diversity was higher in in Coleoptera in organic

fields (P = 0.01) and Diptera and Hemiptera in conventional fields (P < 0.001 in

both), For Araneae we did not find statistical differences (P = 0.21). However,

excluding dominant groups, organic fields showed higher diversity (respectively

for organic and conventional fields: 2.7 ± 0.4 and 2.5 ± 0.3, P = 0.01). The

99

differences between diversity indices calculated including and excluding the

dominant groups were higher for organic (0.8 ± 0.5) than for conventional fields

(0.3 ± 0.4) (P < 0.001 in both cases).

The total estimated biomass of arthropods collected was slightly higher in

organic fields than in conventional fields, although the difference was not

significant (Table 5). By orders, only Collembola showed higher biomass in organic

fields (P < 0.001), and Thysanoptera in conventional fields (P = 0.002). We

searched for factors affecting arthropod biomass through GLMM. As weed richness

was highly correlated with weed abundance (R = 0.72, P < 0.001) and weed

diversity (R = 0.65, P < 0.001), weed richness was discarded from the plausible

factors in GLMM, which included management type, weed abundance, weed

diversity and weed cover as fixed factors, and field pair as random factor

(Appendix D). Three models could be considered candidate models according to

their differences in AICc (<2). The variables included in the best model were

management type, weed abundance, and weed diversity (Table 6), with 27.1% of

the deviance explained (Appendix D). Model 1 differed from model 2 (P = 0.03),

which also included weed cover (27.3% of the deviance explained). Model 3

included the interaction between management and weed diversity (27.3% of the

deviance explained).

Table 6. Parameter estimates from the Generalized Linear Mixed Model with management

(MA), weed abundance (WA) and weed diversity (WD) as factors affecting arthropod biomass.

Parameter Estimate SE P

MA 4.091 0.172 < 0.001

WA 2.036 0.065 < 0.001

WD 3.153 0.139 < 0.001

100

DISCUSSION

In the dryland cereal agroecosystem studied, the first effect of organic farming was

on the weeds, with knock-on effects (Hawes et al., 2003) on the arthropods

community, associated directly with this resource. Finally, the competition with

weeds led into a decreased cereal production, as suggested by the lower number of

cereal plants. The positive effect of a reduction in agrochemical applications on

weed density has been experimentally demonstrated (e.g., Frampton and Dorne,

2007; Hyvönen and Salonen, 2002; Kleijn et al., 2006). Weed and arthropod

communities were also richer in organic fields and, in the case of weeds, more

diverse than those of conventional fields. The average increases in weed

abundance (202%), richness (176%), diversity (133%) and cover (421%) in

organic fields were somewhat higher than those recorded in a dryland cereal area

in northern Spain (Caballero-López et al., 2010; José-María et al., 2010; Romero et

al., 2008), and considerably higher than those reported for studies carried out at

northern latitudes (e.g., Bengtsson et al., 2005; Hole et al., 2005; Moreby et al.,

1994). The higher development of weeds in absence of agrochemical treatment in

these Spanish studies as compared to studies carried out at northern latitudes

might be explained by several facts. First, the weed flora is more diverse in

Mediterranean latitudes (Araújo et al., 2007; Cowling et al., 1996; Thompson,

2005). Second, in most Spain cereal is grown in a traditional two-year rotation

system that creates a mosaic of ploughed, cereal and stubble patches, with some

fallow fields left untilled for several years. Such system allows uncultivated fields

to act as weed reservoirs from which their seeds may easily disperse, building up a

rich weed community in organic cereal fields. In the more intensively cultivated

cereal farmland in northern countries, these uncultivated weed reservoirs are less

frequent, and thus the weed development in organic fields less marked. Third, in

our study area fields are small (less than 2 ha) and field boundaries are narrow

(mean width = 35 ± 25 cm, mean height = 40 ± 23 cm, n = 50, own data), favoring

an easy exchange of seeds and arthropods among fields.

A limitation of our study could be that sampling was restricted to a single

year of organic farming. However, rather than looking at an equilibrium situation,

we were interested in knowing whether a quick response to organic treatment

101

could be observed. Some authors have noticed that rapid positive responses to

agri-environmental measures would imply less costs, and that if an agri-

environmental measure needs several years to become effective, perhaps it should

not be implemented (e.g., Hole et al., 2005). Moreover, the temperature and

precipitation values of the study year were within half a standard deviation of the

average for the last 30 years, suggesting that the results were probably not

influenced by weather conditions. Finally, instead of performing several samplings

through the spring, we restricted our sampling to just one time during May, due to

the relatively short vegetative period in our study area. The sampling dates were

selected to maximize the probability of collecting most weeds and arthropods,

which in our study area have very short life cycles as compared to more northern

latitudes. Besides, sampling effort for plants was adequate, since we sampled 80%

and 89.8% of the species predicted by Clench equation, respectively in organic and

conventional fields (Jiménez-Valverde et al., 2003; Moreno and Halffter, 2001).

As in the study of Romero et al., (2008), in our area Lolium rigidum (Gaudin)

was the only dominant weed in conventional fields, due to its particular resistance

to herbicides (Heap, 1997), and Avena sterilis (L.) and Bromus diandrus (Roth)

were also relatively resistant. When herbicides were suppressed, a more complex

weed community developed, and the prevalence of L. rigidum (Gaudin) decreased

significantly, leaving space to other weeds, particularly broad-leaved species which

are less resistant to the herbicides used (Kudsk and Streibig, 2003). Among these,

several leguminous species were particularly important, since they contribute to

nitrogen fixation, and thus to the development of a richer biocenosys. These

species were Vicia sativa (L.), V. spp., Trifolium angustifolium (L.) y Ornithopus

compressus (L.), which together comprised ca. 7% of weeds in organic fields, as

compared to only 0.2% in conventional fields. Some legumes are also related to

increases in some arthropod groups as flower-consumers, chewing-herbivores and

saprophages (Caballero-López et al., 2010).

The best models selected by the GLMMs showed an influence of

management type and cereal cover on weed richness, as well as an interaction

between both variables. This means that as cereal cover decreased, the richness of

the weed community increased, but only in the sample of organically managed

102

fields. Such relationship was not observed in conventional fields where herbicide

treatment kept weeds under control. On average, organic farming implied a 24%

reduction in the number of cereal plants. Assuming plant numbers are correlated

with total cereal crop, organic farming also determined a similar decrease in

agricultural production. Such a decrease is slightly higher than the 16.5% reported

as mean variation among years in winter cereal production in Spain (MMAMRM,

2010).

As for arthropods, their abundance increased in organic fields compared to

conventional fields (41%). Such increase is similar to those reported in previous

studies (Bengtsson et al., 2005; Frampton and Dorne, 2007; Hole et al., 2005). The

Collembola, Chloropidae (Diptera), and Aphididae (Hemiptera) were found to be

dominant groups. These species were ca. 20% more abundant in organic fields

than in conventional fields, concluding that their proliferation could be a direct

consequence of the farming system. Clough et al., (2007b) also found some

dominant species of the Staphylinidae (Coleoptera) and Moreby et al., (1994) found

an increase of Diptera and Aphids (Hemiptera), the same orders identified as

dominant in the present study. Their higher abundance and proliferation in

organic fields could probably be favored by the greater cover in these fields of

insect-pollinated weeds, particularly those with flowers, the typical niche of most

of these insects. Arthropod richness was a 6.4% higher in organic fields. Most other

studies have also recorded richness increases in organically managed fields

(Clough et al., 2007a; Hadjicharalampous et al., 2002; Hole et al., 2005; Pfiffner and

Niggli, 1996), and the impact of organic management on arthropods has been

interpreted as an indirect result of the impact of agro-chemical suppression on the

vegetation (Siemann et al., 1998). Finally, multivariate models showed that

arthropod biomass was significantly influenced by farming practices, weed

abundance and weed diversity. The best model explained only a 27.1% of the total

deviance, which suggests that additional variables such as landscape complexity,

distance to nearby organic fields, and field size could also be relevant (Clough et al.,

2007a; Concepción et al., 2008).The lower arthropod diversity in organic fields is

explained by the marked dominance in these fields of a few taxa, mainly

Collembola, Chloropidae (Diptera), and Aphididae (Hemiptera). As argued by Shah

et al., (2003), who also found a higher diversity in conventional fields, the

103

Shannon-Wiener diversity index, despite its wide use in biodiversity studies, is

particularly sensitive to changes in the abundance of dominant species in a sample.

In their study, the diversity decrease in organic fields was due to the abundance of

a dominant carabid, Pterostichus melanarius (Illiger). Several other studies also

showed that organic management systems increased arthropod abundance and

richness but not diversity (Booij, 1994; Clark, 1999; Hokkanen and Holopainen,

1986; Kromp, 1999). In our study, the greater abundance in organic fields of the

three dominant groups mentioned above was probably related to a higher

development of the weeds canopy, since Chloropidae adults are flower-consumers

and chewing-herbivores, and Aphididae are suction-herbivores (Caballero et al.,

2010). Without considering these dominant groups, the frequency distribution of

the remaining species indicated a significantly higher diversity in organic fields.

This was consistent with richness values, which were higher in organic than in

conventional fields.

Overall, our results confirm findings from previous studies, and suggest that

organic farming may contribute to preserve biodiversity in the dryland cereal

agroecosystem of our study area. Organic farming could thus be used as a way to

minimize the negative impacts of agricultural intensification, and particularly to

improve habitat quality for many vertebrate consumers such as several

endangered steppe birds inhabiting dry cereal farmland in the Mediterranean

region.

ACKNOWLEDGEMENTS

We thank L. M. Bautista for his help during field work, N. Panizo and R. Sansegundo

for their collaboration in weed and arthropod identification, and A. Torres and J.

Seoane for advice during statistical analyses. Three anonymous reviewers, the

editor and the editor in chief improved the manuscript with their comments. We

also thank all farmers participating in this study. Compensatory payments to

farmers were financed through a project CSIC-HENARSA which aims to develop an

agri-environmental scheme to enhance farmland bird populations in the SPA 139.

Additional funding was provided by projects CGL2005-04893 and CGL2008-02567

104

of the Dirección General de Investigación of the Spanish Ministry for Science and

Innovation.

REFERENCES

Araújo, M.B., Lobo, J.M., Moreno, J.C., 2007.The performance of Iberian protected

areas to conserve terrestrial biodiversity. Conserv. Biol. 21 (6), 1423–1432.

Barrientos, R., Bolonio, L., 2009. The presence of rabbits adjacent to roads

increases polecat road mortality. Biodivers. Conserv. 18, 405–418.

Bates, D., Maechler, M., 2010. lme4: Linear mixed-effects models using S4 classes. R

package version 0.999375-37. http://CRAN.R-project.org/package=lme4

Beecher, N.A., Johnson, R.J., Brandle, J.R., Case, R.M., Young, L.J., 2002. Agroecology

of birds in organic and nonorganic farmland. Conserv. Biol. 16, 1620–1631.

Bengtsson, J., Ahnström, J., Weibull, J., 2005. The effects of organic agriculture on

biodiversity and abundance: a meta-analysis. J. Appl. Ecol. 42, 261–269.

Berger, W.H., Parker, F.L., 1970. Diversity of planktonic foraminifera in deep-sea

sediments. Science. 168, 1345–134.

Biaggini, M., Consorti, R., Dapporto, L., Dellacasa, M., Paggetti, E., Corti, C., 2007. The

taxonomic level order as a possible tool for rapid assessment of arthropod

diversity in agricultural landscapes. Agric. Ecosyst. Environ. 122, 183–191.

Booij, K., 1994. Diversity patterns in carabid assemblages in relation to crops and

farming systems, in: Desender, K., Dufrêne, M., Loreau, M., Luff, M.L.,

Maelfait, J.P. (Eds.), Carabid beetles: Ecology and Evolution. Kluwer

Academic Publishers, Dordrecht, pp. 425–431.

Burnham, K.P., Anderson, D.R. 2002. Model selection and multimodel inference.

Springer, New York.

105

Caballero-López, B., Blanco-Moreno, J.M., Pérez, N., Pujade-Villar, J., Ventura, D.,

Oliva, F., Sans, F.X., 2010. A functional approach to assessing plant–

arthropod interaction in winter wheat. Agr. Ecosyst. Environ. 137, 288–293.

Caruso, T., Pigino, G., Bernin, F., Bargagli, R., Migliorini, M., 2007. The Berger–

Parker index as an effective tool for monitoring the biodiversity of

disturbed soils: a case study on Mediterranean oribatid (Acari: Oribatida)

assemblages. Biodivers. Conserv. 16, 3277–3285.

Chamberlain, D.E., Joys, A., Johnson, P.J., Norton, L., Feber, R.E., Fuller, R.J., 2010

Does organic farming benefit farmland birds in winter? Biol. Letters, 6, 82–

84.

Clark, M.S., 1999. Ground beetle abundance and community composition in

conventional and organic tomato systems of California´s Central Valley.

Appl. Soil. Ecol. 11, 199–206.

Clere, E., Bretagnolle, V., 2001. Disponibilité alimentaire pour les oiseaux en milieu

agricole: biomasse et diversité des arthropodes capturés par la méthode des

pots-pièges. Rev. Ecol-Terre Vie. 56, 275–297.

Clough, Y., Holzschuh, A., Gabriel, D., Purtauf, T., Kleijn, D., Kruess, A., Steffan-

Dewenter, I., Tscharntke, T., 2007a. Alpha and beta diversity of arthropods

and plants in organically and conventionally managed wheat fields. J. Appl.

Ecol. 44, 804–812.

Clough, Y., Kruess, A., Kleijn, D., Tscharntke, T., 2005. Spider diversity in cereal

fields: comparing factors at local, landscape and regional scales. J. Biogeogr.,

32, 2007–2014.

Clough, Y., Kruess, A., Tscharntke, T., 2007b. Organic versus conventional arable

farming systems: Functional grouping helps understand staphylinid

response. Agr. Ecosyst. Environ. 118, 285–290.

Concepción, E.D., Díaz, M., Baquero, R.A., 2008. Effects of landscape complexity on

the ecological effectiveness of agri-environment schemes. Landscape Ecol.

23, 135–148.

106

Colwell, R.K., 2009. EstimateS: Statistical estimation of species richness and shared

species from samples. Version 8.2. User's Guide and application published

at: http://purl.oclc.org/estimates.

Costa, M., Morla, C., Sainz, H., 2004. Los bosques ibéricos: una interpretación

geobotánica. Planeta, Barcelona.

Cotes, B., Campos, M., Pascual, F., Pedro, A. and Ruano, F., 2010. Comparing

taxonomic levels of epigeal insects under different farming systems in

Andalusian olive agroecosystems. Appl. Soil Ecol. 44, 228–236.

Cowling, R.M., Rundel, P.W., Lamont, B.B., Arroyo, M.K., Arianoutsou, M., 1996.

Plant diversity in mediterranean-climate regions. Trends Ecol. Evol. 11 (9),

362–366.

Diekötter, T., Wamser, S., Wolters, V., Birkhofer, K., 2010. Landscape and

management effects on structure and function of soil arthropod

communities in winter wheat. Agr. Ecosyst. Environ., 137 (1-2), 108–112.

Donald, P.F., Sanderson, F.J., Burfield, I.J., Van Bommel, F.P.J., 2006. Further

evidence of continent-wide impacts of agricultural intensification on

European farmland birds. Agr. Ecosyst. Environ. 116, 189–196.

EEA, 2004. High Nature Value Farmland – Characteristics, Trends and Policy

Challenges. European Environment Agency, Copenhagen.

Fauvel, G., 1999. Diversity of Heteroptera in agrosystems: role of sustainability and

bioindication. Agr. Ecosyst. Environ. 74, 275–303.

Firbank, L.G., 2009. It’s not enough to develop agriculture that minimizes

environmental impact. Int. J. Agr. Sustain. 7, 1–2.

Foley, J.A., DeFries, R., Asner, G.P., Barford, C., Bonan, G., Carpenter, S.R., Chapin,

F.S., Coe, M.T., Daily, G.C., Gibbs, H.K., Helkowski, J.H., Holloway, T., Howard,

E.A., Kucharik, C.J., Monfreda, C., Patz, J.A., Prentice, I.C., Ramankutty, N.,

Snyder, P.K., 2005. Global consequences of land use. Science. 309, 570–574.

107

Frampton, G.K., Dorne, J.L.C.M., 2007. The effects on terrestrial invertebrates of

reducing pesticide inputs in arable crop edges: a meta-analysis. J. Appl. Ecol.

44, 362–373.

Fuller, R.J., Norton, L.R., Feber, R.E., Johnson, P.J., Chamberlain, D.E., Joys, A.C.,

Mathews, F., Stuart, R.C., Townsend, M.C., Manley, W.J., Wolfe, M.S.,

Macdonald, D.W., Firbank, L.G., 2005. Benefits of organic farming to

biodiversity vary among taxa. Biol. Letters, 1, 431–434.

Gabriel, D., Roschewitz, I., Tscharntke, T., Thies, C., 2006. Beta diversity at different

spatial scales: Plant communities in organic and conventional agriculture.

Ecol. Appl., 16, 2011–2021.

Gabriel, D., Sait, S.M., Hodgson, J.A., Schmutz, U., Kunin, W.E., Benton, T.G., 2010.

Scale matters: the impact of organic farming on biodiversity at different

spatial scales. Ecol. Lett., 13, 858–869.

Geiger, F., Bengtsson, J., Berendse, F., Weisser, W.W., Emmerson, M., Morales, M.B.,

Ceryngier, P., Liira, J., Tscharntke, T., Winqvist, C., Eggers, S., Bommarco, R.,

Pärt, T., Bretagnolle, V., Plantegenest, M., Clement, L.W., Dennis, C., Palmer,

C., Oñate, J.J., Guerrero, I., Hawro, V., Aavik, T., Thies, C., Flohre, A., Hänke, S.,

Fischer, C., Goedhart, P.W., Inchausti, P., 2010. Persistent negative effects of

pesticides on biodiversity and biological control potential on European

farmland. Basic Appl. Ecol. 11 (2), 97–105.

Gibson, R.H., Pearce, S., Morris, R.J., Symondson, W.O.C., Memmott, J., 2007. Plant

diversity and land use under organic and conventional agriculture: a whole-

farm approach. J. Appl. Ecol., 44, 792–803.

Hadjicharalampous, E., Kalburtji, K.L., Mamolos, A.P., 2002. Soil arthropods

(Coleoptera, Isopoda) in organic and conventional agroecosystems. Environ.

Manage. 29 (5), 683–690.

Hawes, C., Haughton, A.J., Osborne, J.L., Roy, D.B., Clark, S.J., Perry, J.N., Rothery, P.,

Bohan, D.A., Brooks, D.R., Champion, G.T., Dewar, A.M., Heard, M.S., Woiwod,

I.P., Daniels, R.E., Young, M.W., Parish, A.M., Scott, R.J., Firbank, L.G. & Squire,

G.R., 2003. Responses of plants and invertebrate trophic groups to

108

contrasting herbicide regimes in the Farm Scale Evaluations of genetically

modified herbicide-tolerant crops. Phil. Trans. R. Soc. Lond. B. 358, 1899–

1913.

Heap, I.M., 1997. The occurrence of herbicide-resistant weeds worldwide. J. Pestic.

Sci. 51 (3), 235–243.

Hódar, J.A., 1996. The use of regression equations for estimation of arthropod

biomass in ecological studies. Acta Oecol. 17, 421–433.

Hokkanen, H., Holopainen, J.K., 1986. Carabid species and activity in biologically

and conventionally managed cabbage fields. J. Appl. Entomol. 102, 353–363.

Hole, D.G., Perkins, A.J., Wilson, J.D., Alexander, I.H., Grice, P.V., Evans, A.D., 2005.

Does organic farming benefit biodiversity? Biol. Conserv. 122 (1), 113–130.

Hyvönen, T., Salonen, J., 2002. Weed species diversity and community composition

in cropping practices at two intensity levels – a six-year experiment. Plant

Ecol. 154, 73–81.

INE, 2009. Climatología. Anuario Estadístico de España.

www.ine.es/jaxi/menu.do?type=pcaxis&path=/t43/a012/a1998&file=pcaxis,

Last access: February 20, 2010.

Jiguet, F., Arroyo, B., Bretagnolle, V., 2000. Lek mating systems: a case study in the

Little Bustard Tetrax tetrax. Behav. Process. 51, 63–82.

Jimenez-Valverde, A., and Hortal, J., 2003. Las curvas de acumulación de especies y

la necesidad de evaluar la calidad de los inventarios biológicos. Rev. Ibér.

Aracnol., 8, 151–161.

José-María, L., Armengot, L., Blanco-Moreno, J.M., Bassa, M., Xavier Sans, F., 2010.

Effects of agricultural intensification on plant diversity in Mediterranean

dryland cereal fields. J. Appl. Ecol., 47, 832–840.

Kleijn, D., Baquero, R.A., Clough, Y., Diaz, M., Esteban, J., Fernandez, F., Gabriel, D.,

Herzog, F., Holzschuh, A., Johl, R., Knop, E., Kruess, A., Marshall, E.J.P.,

Steffan-Dewenter, I., Tscharntke, T., Verhulst, J., West, T.M., Yela, J.L., 2006.

109

Mixed biodiversity benefits of agri-environment schemes in five European

countries. Ecol. Lett. 9, 243–254.

Kromp, B., 1999. Carabid beetles in sustainable agriculture: a review on pest

control efficacy, cultivation impacts and enhancement. Agr. Ecosyst.

Environ. 40, 71–93.

Kudsk, P., Streibig, J.C., 2003. Herbicides: a two-edged sword. Weed Res. 43 (2),

90–102.

Littel, R.C., Milligen, G.A., Stroup, W.W., Wolfinger, R.D., Schabenberger, O., 2006.

SAS for Mixed Models, 2nd edn. SAS Institute Inc., Cary, NC: SAS Institute.

Lundkvist, A., Salomonsson, L., Karlsson, L., Gustavsson, A.M.D., 2008. Effects of

organic farming on weed flora composition in a long term perspective. Eur.

J. Agron. 28, 570–578.

Millennium Ecosystem Assessment, 2005. Ecosystems and Human Well-being:

Biodiversity Synthesis. World Resources Institute, Washington, DC.

MMAMRM, 2010. Balances de Gestión de Cereales.

www.mapa.es/es/agricultura/pags/cereales/Nacional/ConjuntoCereales.htm,

Last access: February 20, 2010.

Moreby, S.J., Aebischer, N.J., Southway, S.E., Sotherton, N.W., 1994. A comparison of

the flora and arthropod fauna of organically and conventionally grown

winter wheat in southern England. Ann. Appl. Biol. 125, 13–27.

Moreno, C.E., Halffter, G., 2001. On the measure of sampling effort used in species

accumulation curves. J. Appl. Ecol., 38, 487–490.

Moya-Laraño, J., Wise, D.H., 2007. Two simple strategies to increase the power of

experiments with multiple response variables. Basic Appl. Ecol. 8, 398–410.

Pfiffner, L., Niggli, U., 1996. Effects of biodynamics, organic and conventional

farming on ground beetles (Col. Carabidae) and other epigeic arthropods in

winter wheat. Biol. Agric. Hortic. 12, 353–364.

110

Piha, M., Tiainen, J., Holopainen, J., Vepsalainen, V., 2007. Effects of land-use and

landscape characteristics on avain diversity and abundance in a boreal

agricultural landscape with organic and conventional farms. Biol. Conserv.,

140, 50–61.

R Development Core Team, 2010. R: A language and environment for statistical

computing. R Foundation for Statistical Computing, Vienna, Austria. www.R-

project.org. Last access: February 10, 2011.

Romero, A., Chamorro, L., Sans, X., 2008. Weed diversity in crop edges and inner

fields of organic and conventional dryland winter cereal crops in NE Spain.

Agr. Ecosyst. Environ. 124, 97–104.

Roschewitz, I., Gabriel, D., Tscharntke, T., Thies, C., 2005. The effects of landscape

complexity on arable weed species diversity in organic and conventional

farming. J. Appl. Ecol., 42, 873–882.

Royal Society, 2009. Reaping the Benefits: Science and the Sustainable

Intensification of Global Agriculture. The Royal Society, London.

Schmidt, M.H., Clough, Y., Schulz, W., Westphalen, A., Tscharntke, T., 2006. Capture

efficiency and preservation attributes of different fluids in pitfall traps. J.

Arachnol. 34, 159–162.

Schmidt, M.H., Roschewitz, I., Thies, C., Tscharntke, T., 2005. Differential effects of

landscape and management on diversity and density of ground-dwelling

farmland spiders. J. Appl. Ecol., 42, 281–87.

Shah, P.A., Brooks, D.R., Ashby, J.E., Perry, J.P., Woiwod, I.P., 2003. Diversity and

abundance of the coleopteran fauna from organic and conventional

management system in southern England. Agr. Forest. Entomol. 5, 51–60.

Siemann, E., Tilman, D., Haarstad, J., Ritchie, M., 1998. Experimental tests of the

dependence of arthropod diversity on plant diversity. Am. Nat. 152, 738–

750.

111

Sunderland, K., Samu, F., 2000. Effects of agricultural diversification on the

abundance, distribution, and pest control potential of spiders: a review.

Entomol. Exp. Appl. 95, 1–13.

Thompson, J.D., 2005. Plant Evolution in the Mediterranean. Oxford University

Press, Oxford.

Walter, H., 1994. Vegetation of the Earth and Ecological Systems of the Geo-

biosphere. third ed. Heidelberg Science Library, Springer-Verlag, Germany.

Weibull, A.C., Östman, Ö., 2003. Species composition in agroecosystems: The effect

of landscape, habitat, and farm management. Basic Appl. Ecol. 4, 349–361.

Weibull, A.C., Ostman, Ö., Granqvist, A., 2003. Species richness in agroecosystems:

the effect of landscape, habitat and farm management. Biod. Conserv., 12,

1335–1355.

Wilson, J.D., Evans, A.D., Grice, P.V., 2009. Bird Conservation and Agriculture.

Cambridge University Press, Cambridge.

Wilson, J.D., Evans, A.D., Grice, P.V., 2010, Bird conservation and agriculture: a

pivotal moment?. Ibis, 152, 176–179.

112

Appendix A. Complete list of weed species identified, ordered by the frequency with which they were

recorded in organically managed and conventional fields. Values are percentages of each species found in

both field types.

Species Organic fields Conventional fields

Lolium rigidum (Gaudin) 50.3 80.2

Galium tricornutum (Dandy) 7.1 1.8

Bromus diandrus (Roth) 6.6 4.2

Anacyclus clavatus (Desf.) 4.2 0.7

Conyza canadensis (L.) 4.0 0.6

Raphanus raphanistrum (L.) 3.8 0.6

Avena sterilis (L.) 3.0 3.3

Vicia sativa (L.) 2.8 0.0

Polygonum aviculare (L.) 2.6 2.9

Filago pyramidata (L.) 1.7 0.1

Trifolium angustifolium (L.) 1.7 0.0

Filago lutescens (Jord.) 1.1 1.2

Vicia spp 0.9 0.1

Picnomon acarna (L.) 0.9 0.8

Lactuca serriola (L.) 0.8 0.2

Ornithopus compressus (L.) 0.7 0.1

Linaria viscosa (L.) 0.7 0.0

Spergula arvensis (L.) 0.7 0.0

Euphorbia serrata (L.) 0.6 0.0

Vicia ervilia (L.) 0.6 0.0

Filago gallica (L.) 0.6 0.0

Convolvulus arvensis (L.) 0.6 1.5

Anagallis arvensis (L.) 0.4 0.3

Carduus tenuiflorus (Curtis) 0.4 0.0

Hordeum murinum (L.) 0.4 0.1

Aegilops geniculata (Roth) 0.2 0.0

Andryala integrifolia (L.) 0.2 0.2

Ranunculus arvensis (L.) 0.2 0.0

Taeniatherum caput-medusae (L.) Nevski 0.2 0.3

Lathyrus sp 0.2 0.0

Anchusa azurea (Mill.) 0.2 0.0

Bromus squarrosus (L.) 0.2 0.0

Adonis aestivalis (L.) 0.1 0.0

Lupinus angustifolius (L.) 0.1 0.1

Centaurea cianus (L.) 0.1 0.0

Chenopodium album (L.) 0.1 0.2

Cnicus benedictus (L.) 0.1 0.1

Papaver rhoeas (L.) 0.1 0.1

Picris echioides (L.) 0.1 0.0

113

Species Organic fields Conventional fields

Torilis nodosa (L.) 0.1 0.0

Trifolium campestre (Schred. in Sturn) 0.1 0.0

Senecio vulgaris (L.) 0.1 0.1

Spergularia rubra (L.) J. Presl & C. Presl 0.1 0.0

Arabidopsis thaliana (L.) Heynh. in Holl & Heynh. 0.0 0.0

Ononis spinosa (L.) 0.0 0.0

Sherardia arvensis (L.) 0.0 0.0

Sonchus oleraceus (L.) 0.0 0.0

Taraxacum officinale (Weber) 0.0 0.0

Amaranthus albus (L.) 0.0 0.2

Cynodon dactylon (L.) Pers. 0.0 0.1

Rumex pulcher (L.) 0.0 0.1

Veronica hederifolia (L.) 0.0 0.2

Appendix B. Results of Generalized Linear Mixed Models (GLMMs) where management (MA) and

cereal cover (CC) were factors affecting weed richness. Field pair was the random factor. The best

models (1 and 2) were determined according to the lowest corrected Akaike's Information

Criterion (AICc) and ANOVA test (P is given, when ΔAICc between one model and the best was less

than two). The percentage of the explained deviance, degrees of freedom (d.f.) and model log-

likelihood (LL) are also given

Model Number AICc AICc Explained deviance d.f. LL P

1 CC + MA + CC*MA 82.3 0.00 56.8 5 -35.6 0.12

2 CC + MA 83.4 1.10 54.6 4 -37.4 < 0.001

3 MA 83.7 1.43 52.9 3 -38.8

4 CC 169.6 87.39 0.8 3 -81.8

Appendix C. Arthropod orders and families identified, and number of individuals collected in

organic and conventional fields.

Order Family Organic fields Conventional fields

Acari Gamasidaea 2550 1639

Oribatidab 23 4

Araneae Anyphaenidae

2

Atypidae 23 12

Ctenizidae

1

Dictynidae 1

Gamasidae 1 1

Linyphiidae 467 530

114

Order Family Organic fields Conventional fields

Lycosidae 66 89

Oonipidae 2 3

Oxyopidae 17 17

Palpimanidae 2 5

Pholcidae

1

Sicariidae

1

Theraphosidae 2

Telemidae 1

Theraphosidae

1

Theridiidae 31 17

Titanoecidae

1

Thomisidae 33 39

Uloboridae

1

Zoridae 77 50

Zoropsidae 61 62

Coleoptera Aesalidae

3

Anthribidae 4

Anthicidae 59 25

Brostrichidae 1

Bruchidae 2

Byrrhidae 1

Cantharidae 73 82

Carabidae 287 305

Cerambycidae 28 8

Chrysomelidae 84 69

Ciidae 1

Coccinelidae 162 153

Curculionidae 119 145

Dermestidae 15 12

Dryopidae 5 1

Elateridae 40 9

Erotylidae

1

Gyrinidae

2

Histeridae

1

Lampyridae 12 9

Malachidae 3 4

Meloidae 1

Nitidulidae 54 56

Omaliinae

2

Scarabeidae 40 19

Scydmaenidae 2 1

Staphylinidae 140 250

Silphidae 204 287

Silvanidae 1 2

Tenebrionidae 2

Trogidae 2 7

115

Order Family Organic fields Conventional fields

Collembolab Collembola 9558 3630

Diplura Campodeidae 9 70

Japygidae 2 1

Diptera Acroceridae 40 28

Anthomiidae 1

Asilidae 41 93

Bibionidae 5 2

Camillidae 253 452

Cecidomyiidae 534 365

Ceratopogonidae 2

Chloropidae 6496 3147

Conopidae 65 28

Culicidae 38 67

Dixidae

1

Fanniidae 1

Heleomyzidae

2

Hippoboscidae 75 49

Lauxaniidae 11 23

Lonchopteridae 2

Milichiidae

1

Muscidae 183 277

Mycethophilidae 2

Otitidae 3 1

Phoridae 82 156

Pipunculidae 229 242

Platystomatidae 19 2

Psilidae 21 40

Ptychopteridae 18 13

Sarcophagidae 2

Scathophagidae 92 105

Scatopsidae 4 2

Scenopinidae

1

Sepsidae

8

Sphaeroceridae 1 3

Stratiomyidae 7 18

Syrphidae 99 237

Tabanidae 3 17

Tachinidae

1

Tethritidae 1

Therevidae

3

Trichoceridae 112 301

Trigonalidae 85 111

Vermileonidae 55 52

Xylophagidae

8

Embioptera Oligotomidae

1

116

Order Family Organic fields Conventional fields

Hemiptera Acanthosomidae 28 5

Alydidae 19 1

Anthocoridae

2

Aphididae 23014 13130

Aphrophoridae 815 1674

Cicadellidae 2

Cicadidae 247 261

Cimicidae 1 6

Delphacidae 2

Lygaeidae 15 5

Miridae 44 1

Nabidae 191 36

Pentatomidae 56 2

Pseudococcidae

1

Psyllidae

6

Reduviidae 12 1

Rhopalidae 6

Scutelleridae 3 1

Hymenoptera Andrenidae 1

Apidae 2

Cynipidae

2

Evaniidae

4

Formicidae 1863 867

Pamphiliidae

1

Pompilidae

2

Sapygidae

4

Siricidae 1

Trichogrammatidae

21

Vespidae

2

Xyelidae

3

Isopoda Philosciidae 2 1

Lepidoptera Papilionidae

3

Pyralidae 2 1

Mecoptera Boreidae

3

Panorpidae 2 1

Miriapodac Diplopodab 4 5

Neuroptera Ascalaphidae 3 1

Hemerobiidae 4 3

Myrmeleonidae 3 2

Odonata Coenagrionidae

1

Opinilionida Phalangiidae 5 7

Orthoptera Acrididae 45 36

117

Order Family Organic fields Conventional fields

Gryllidae 56 13

Pamphagidae

2

Tettigoniidae 25 20

Trydactylidae 1 4

Gryllotalpidae 30 19

Psocoptera Psocidae

2

Siphonaptera Hystrichopsyllidae 109 174

Thysanoptera Thripidae 1015 2510

Thysanura Lepismatidae 10 1

Total

50488 32334

a SubOrder

b Class

c SubPhylum

Appendix D. Results of Generalized Linear Mixed Models (GLMMs) where management (MA), weed

abundance (WA), weed diversity (WD), and weed cover (WC) were factors affecting arthropod

biomass. Field pair was the random factor. The best model was determined according to the lowest

corrected Akaike's Information Criterion (AICc) and ANOVA test (P is given, when ΔAIC between

one model and the best was less than two). The percentage of the explained deviance, degrees of

freedom (d.f.) and model log-likelihood (LL) are also given.

Model number AICc AICc Explained deviance

d.f. LL P

1 MA + WA +WD 1684.6 0 27.1 5 -834.7 0.03

2 MA + WA +WD + WC 1685.6 0.98 27.3 6 -833.0 1

3 MA + WA +WD + MA*WD 1685.6 0.98 27.3 6 -833.2

4 MA + WA +WD + WC + MA*WD 1687.0 2.40 27.4 7 -831.5

5 MA + WD + MA*WD 1715.6 31.00 25.6 5 -849.9

6 MA + WA + MA*WA 1785.6 101.00 21.1 5 -885.1

7 MA + WA 1802.0 117.40 22.1 4 -894.8

118

119

CAPÍTULO 4

120

Este capítulo reproduce íntegramente el siguiente artículo:

Ponce, C., Bravo, C., & Alonso, J.C. 2014. Effects of agri-environmental schemes on

farmland birds: Do food availability measurements improve patterns obtained

from simple habitat models?. Ecology and Evolution 4: 2834–2847.

121

CAPÍTULO 4

Effects of agri-environmental schemes on

farmland birds: do food availability

measurements improve patterns obtained

from simple habitat models?

Carlos Poncea, Carolina Bravoa, Juan Carlos Alonsoa

a Dep. Ecología Evolutiva, Museo Nacional de Ciencias Naturales, CSIC, José

Gutiérrez Abascal 2, E-28006 Madrid, Spain

122

ABSTRACT

Studies evaluating agri-environmental schemes (AES) usually focus on responses of single

species or functional groups. Analyses are generally based on simple habitat

measurements but ignore food availability and other important factors. This can limit our

understanding of the ultimate causes determining the reactions of birds to AES. We

investigated these issues in detail and throughout the main seasons of a bird’s annual cycle

(mating, post-fledging and wintering) in a dry cereal farmland in a Special Protection Area

for farmland birds in central Spain. First, we modeled four bird response parameters

(abundance, species richness, diversity and ‘Species of European Conservation Concern’

[SPEC]-score), using detailed food availability and vegetation structure measurements

(food models). Second, we fitted new models, built using only substrate composition

variables (habitat models). Whereas habitat models revealed that both, fields included and

not included in the AES benefited birds, food models went a step further and included seed

and arthropod biomass as important predictors, respectively in winter and during the

postfledging season. The validation process showed that food models were on average

13% better (up to 20% in some variables) in predicting bird responses. However, the cost

of obtaining data for food models was five times higher than for habitat models. This novel

approach highlighted the importance of food availability-related causal processes involved

in bird responses to AES, which remained undetected when using conventional substrate

composition assessment models. Despite their higher costs, measurements of food

availability add important details to interpret the reactions of the bird community to AES

interventions, and thus facilitate evaluating the real efficiency of AES programs.

Keywords

Agri-environmental scheme, agricultural intensification, biomass, habitat

management, steppe birds, wildlife conservation.

123

INTRODUCTION

The demand of more food and biofuel (Tilman et al. 2011; Miyake et al. 2012) from

modern agricultural activities has caused the decline of many species inhabiting

farmland areas (Donald et al. 2001). Increased use of chemicals (pesticides,

fertilizers, etc.), loss of noncropped habitats and loss of crop diversity are some of

the most important factors affecting plant and animal populations in these

ecosystems (Chamberlain et al. 2000; Vickery et al. 2001; Robinson and Sutherland

2002; Benton et al. 2003). However, agri-environmental schemes (AES, hereafter)

are intended to reverse the environmental impacts of modern farming techniques

on biodiversity (Stoate et al. 2009). It is generally accepted that an increase in

habitat heterogeneity has a positive influence on biodiversity (Wuczyński et al.

2011). The European Union and the United States of America have spent several

billion dollars in AES programs (Kleijn et al. 2006; Gabriel et al. 2010), but their

effectiveness is still somehow questioned because different studies have reported

contradictory results (Tscharntke et al. 2005; Kleijn et al. 2006). These differences

may have been due to differences in the scale of study, with most clearly positive

effects at local scales (Perkins et al. 2011) compared with larger scales (Verhulst et

al. 2007; Davey et al. 2010), or in studies designed to enhance certain declining

species (Wilson et al. 2009; Kleijn et al. 2011). AES are usually implemented at

field scale, without controlling for the spatial complexity (vegetation structure and

substrate diversity around managed fields) that affects the variables under study

(Kleijn and Sutherland 2003; Tscharntke et al. 2005; Concepción et al. 2008;

Gabriel et al. 2010; Winqvist et al. 2011). AES design and the research on their

effectiveness usually focus on responses of just one or a few species (Breeuwer et

al. 2009), although other species (MacDonald et al. 2012) or functional groups

(granivorous, insectivorous, etc.; Henderson et al. 2000) may also obtain benefits.

AES studies rarely analyse the responses of the whole bird community, ignoring

that biodiversity maintenance should be a priority (Ekroos et al. 2014). Finally,

although habitat and feeding requirements of species change through the year

(Marfil-Daza et al. 2013), most AES studies evaluate effectiveness during a single

season.

124

In the present study, we investigated the effects of an AES on a steppe bird

community in a dry cereal farmland area in central Spain. We did this by analyzing

various abundance and diversity parameters, which define direct bird responses of

the farmland bird community to the AES (Díaz et al. 2006). The populations of 17

dry farmland bird species present in Spain are rapidly declining (Escandell et al.

2011), even faster here than in other European countries (EBCC 2010). Despite

this, Spain still holds the most important breeding populations of several species

classified as endangered at a continentwide scale, for example, the pin-tailed

sandgrouse (Pterocles alchata), lesser kestrel (Falco naumanni) or great bustard

(Otis tarda). Also, Spain holds significant wintering populations of common

European species like the meadow pipit (Anthus pratensis) or the skylark (Alauda

arvensis). Thus, Spain has the highest impact on the European Farmland Bird Index

(EFBI), an indicator for biodiversity health on farmland areas (Butler et al.

2010a,b). As the scheme prescriptions and measures of our AES (Table 1) were

designed for a broad spectrum of bird species, the results of this study may be

considered as widely applicable for managers and ecologists in general.

125

Table 1. List of field types and prescriptions of the agri-environmental scheme (AES) implemented in the study area

Field type Short name Origin Scheme prescriptions

Legume LegAES AES Growing organic legumes. Sowing Seed on ploughed fields (190kg/ha) in October with a mixture of up to 20% cereal seed. No use of dressed seed. No agricultural activities (weed and arthropod control, tillage tasks, fertilizer applications) from sowing to harvest after 10 July1

LegNAT Natural, non-AES farming Not applicable

Cereal stubble CerStubbleAES AES Maintenance of cereal stubble. No agricultural activities from usual harvest date from June-July to 31 December and from 1 April to 1 July. Tillage tasks allowed from 1st January to 31st March without use of herbicides or insecticides

CerStubbleNAT Natural, non-AES farming Not applicable

Fallow FallowAES AES Interruption of the cereal production for > 1 years. No agricultural activities (weed and arthropod control, tillage tasks, fertilizer applications) are allowed from July to the next July. The agreement can be renewed annually. Fields included in this AES must have been used for agricultural purposes on the last three years

FallowNAT Natural, non-AES farming Not applicable

Legume stubble LegStubbleAES AES No scheme prescriptions. It comes from the LegAES measure

LegStubbleNAT Natural, non-AES farming Not applicable

Cereal crop CerNAT Natural, non-AES farming Not applicable

Plough Plough Natural, non-AES farming Not applicable

Plough with sprouted weeds

Plough2 Natural, non-AES farming Not applicable

Edge Edge Natural, non-AES farming Not applicable

1 Although AES limit the harvest date from 10 July, we accepted farmers harvesting legume fields earlier (but always after 31 June) to feed sheep or to collect

the seed for the following sowing season

126

The AES was funded by a program of preventive, corrective, and

compensatory measures to balance the impact of the M-50 and R-2 highways on

the population of great bustards and other steppe-land birds in the Important Bird

Area (IBA) Talamanca- Camarma and the Site of Community Importance cuenca de

los ríos Jarama y Henares. The two highways were built in the inner border of the

Special Protection Area (SPA) 139 Estepas cerealistas de los ríos Jarama y Henares,

which is included in the IBA. To implement the AES, first we contacted with

agricultural agents to prepare a meeting with farmers from the SPA. Second, we

evaluated all fields farmers offered to be included in the AES following suitability

criteria such as distance to power lines, fences, towns, etc. Third, once a field was

accepted and AES measure implemented on it, we made periodic checks of each

field. Finally, payments to farmers were performed by the company operating the

highways.

However, our main purpose was not just to test the AES effectiveness. Our

first major objective was to explore the effects of a differing quality of the field data

commonly used to investigate bird responses to AES. We did this by comparing the

predictive capacity of response models based on simple habitat measurements

(called habitat models hereafter) with that of models also based on habitat

variables and much more detailed food availability and vegetation structure

measurements (food models hereafter). We also analyzed the cost in terms of

money and effort spent on each set of models, because it is expected that the

amount of time and funds invested should correspond to the quality of the results

obtained. Most AES effectiveness studies have been carried out with relatively low

investment in field work, often using the composition of substrates selected by the

birds before and after AES implementation (see review in Kleijn and Sutherland

2003). The only conclusion that can be drawn from these studies is the positive,

neutral or negative effect of the AES on individual behavior, whereas in most cases

the ultimate causes, processes and population level consequences remain largely

unknown. However, it has been suggested that the association between bird

species and their habitat is determined by the quantity and quality of the resources

provided (functional space available to a species), not only the habitat per se

(Boyce and McDonald 1999; Butler and Norris 2013). For example, it is currently

admitted that agriculture intensification has determined massive declines of

127

farmland bird species, and these declines are due to different processes as reduced

food resources (food availability hypothesis, Newton 2004), reduced refuge quality

(refuge and nesting hypothesis; Benton et al. 2003), or both in some cases

(Campbell et al. 1997; Butler et al. 2007, 2010a,b). Some studies on these questions

are species-specific and do not consider the whole bird community (Breeuwer et

al. 2009; Bretagnolle et al. 2011). However, it is considered that biodiversity loss

can compromise many ecosystem services (Cardinale et al. 2012) and the impacts

of species loss on primary productivity are comparable with impacts from climate

warming (Hooper et al. 2012; Tilman et al. 2012).

Our second main objective was to compare different bird responses:

abundance, richness, diversity and SPECscore, an index based on the Species of

European Conservation Concern (BirdLife International 2004) among three

periods of the annual cycle: wintering, mating, and postfledging. Most studies have

explored ways to enhance breeding success, but very few have been carried out

during the wintering or postfledging seasons, which are also critical for birds. It is

known that ecological circumstances during the nonbreeding seasons

(postfledging and wintering) may affect body condition and survival rates

(Siriwardena et al. 2000; Stoate et al. 2004), and influence the dynamics of the

population (Siriwardena et al. 2007; Butler et al. 2010a,b). Comparisons among

seasonal models enabled us to investigate in detail the processes involved in bird

responses to AES, that is, which aspects of birds’ requirements are better fulfilled

by the agri-environmental measures, and which part of their annual cycle is more

influenced by these measures. To our knowledge, this is the first study comparing

the predictive power of models using field variables of differing quality and

exploring the responses of the whole bird community in different seasons.

Our main study hypotheses were that (i) as birds requirements differ

throughout different periods of their annual cycle, agri-environmental schemes

can lead to different effects in different seasons, and (ii) improving food

availability measurements should lead to significantly higher predictive power

than just using simple habitat measurements, which sometimes may compensate

for the higher field work costs incurred.

128

MATERIAL AND METHODS

Study area, field types, and agri-environmental measures

The field work was carried out in the SPA 139 Estepas cerealistas de los ríos

Jarama y Henares, located in Madrid region (central Spain), where an AES has been

running since 2003. Specifically, we sampled four sites within this area (Fig. 1).

The region has dry cereal cultivation as its main land use, and all the sites share

major environmental–climatic, biogeographic conditions, as well as a similar

steppe bird community.

Figure 1. Location of the study area in the Iberian Peninsula. The figure at the right shows the SPA

139 Estepas Cerealistas de los Ríos Jarama y Henares and the four sites (ellipses) where field work

was carried out.

The study area was a typical non-intensive farmland area, with small fields,

margins between neighbour and the presence of legumes and cereal stubbles

(managed differently from the AES, with the use of pesticides) and fallow fields.

Most cereal was grown following a traditional 2-year rotation system (fields are

cultivated every second year). We mapped habitat types on GIS-based maps along

the transects (200 X 500 m) to calculate the surface of land uses on the same day

when bird censuses were carried out. We defined eight field types (Table 1). The

agri-environmental measures implemented were maintenance of cereal stubble,

growing legumes organically (vetch Vicia sativa), and interruption of the cereal

129

production (fallows; Table 1). The managed surface for transects where birds were

censused ranged from 0% to 76%. We measured the width of 50 margins at

random locations and considered this measure constant throughout the study

(Lane et al. 1999).

Bird surveys

We censused birds from 2006 to 2009 in the winter (between 15 and 31

December), mating (between late April and early May) and postfledging seasons

(early July). We carried out one census per season and site (four censuses in each

season), which is enough to get reliable information about habitat selection and

the above-mentioned parameters (Hanspach et al. 2011). The observer (CP)

walked along 9–14 linear transects per site, each 500 m long and 100 m wide at

each side of the path (totaling 660 transects (220 in each of the three periods –

wintering, mating and post-fledging–) and more than 24,000 birds (see Appendix

S1 in Supporting information). Within each site, a distance of 100 m was kept

between the end of a transect and the beginning of the next transect. Transects

within each site were located at paths that are only used by sheep and farmers. The

path within each site was circular and the observer avoided double counting by

not considering birds near the beginning and end of this path and looking where

flying birds landed during each survey. There was no spatial correlation between

consecutive transects in a site (Appendix S2). Birds flying over transects were only

taken into account if they were clearly using the field (hunting, hovering, etc.). For

each bird flock spotted, we recorded its species composition, number of

individuals, and field type where it landed.

Food availability

Abundance of arthropods, seeds, and vegetation were sampled in three fields of

each substrate type, in the four study sites in the SPA from spring 2006 to spring

2008 (three mating seasons, two postfledging seasons and two wintering seasons)

to obtain average representative values for each substrate in each season and site.

Arthropods were sampled visually following the same methodology as Lane et al.

(1999). Two observers walked at low speed (120 m/h, to avoid effects of

130

detectability due to vegetation characteristics) along linear transects (30 m long

and 0.5 m wide at both sides) in the middle of each field. We counted and identified

every specimen to the most accurate taxonomic level possible by means of visible

characters (Oliver and Beattie 1993). We collected a random sample of 7515

arthropods, 12% of the those detected) and assessed the mean biomass of each

group using linear regression of weight as function of body length, which was

estimated for each individual during the transect (Hódar 1996; Ponce et al. 2011).

We measured all collected arthropods in the laboratory. The mean lengths

obtained allowed us to estimate the length of the observed and noncollected

specimens and to assess the biomass. Finally, for each site, transect, and season, we

calculated the surface of each field type and so obtained a weighted average

biomass per hectare.

Vegetation abundance and structure were assessed throwing a metal

square (25 X 25 cm) at 20 random locations in the middle of each field. Data

recorded were total plant cover (%) and height (minimum, maximum, mean and

most frequent, in cm). In these samples, we evaluated the roughness of the ground

(the degree of flatness of each field) using three categories (low, medium, or high).

We also estimated visually the total numbers of seeds on plants and on the ground

classifying them in four size categories based on their maximum length (<1 mm, 1–

5 mm, 5–10 mm, and >10 mm). We spent the time necessary to count each seed,

regardless of the vegetation present or the field type. We collected a sample of

seeds to measure, weigh, and estimate the mean biomass for each size in the

laboratory. Finally, for each site and season, we calculated the average seed

biomass in each substrate type.

The costs for habitat and food models were calculated according to the

money and time required to obtain the data. We considered field and laboratory

work for food models but only field work for habitat models. Money spent in food

models included costs of sampling and weighing seeds and arthropods, measuring

vegetation structure and recording field uses. In the case of habitat models, the

costs were those of mapping habitat land uses. Both models shared identical bird

census costs. We estimated the costs considering salaries, travel, and daily

subsistence allowances of observers, and field and laboratory material (excluding

131

those provided by the research center). Time (h) needed for each set of models

was calculated according to the effort required for obtaining field and laboratory

data.

Statistical analyses

The response variables were total bird abundance, richness (number of species),

Shannon–Wiener diversity index and SPEC-score in each transect, season, and site.

The SPEC (Species of European Conservation Concern) is the conservation status

of all wild birds in Europe (BirdLife International 2004). The SPEC-score index is

important because it gives more importance to species of major concern in Europe.

Bird species are classified into five categories (from SPEC 1 to SPEC 3 plus Non-

SPEC and Non-SPECE (previously SPEC 4, not detected during the surveys). For the

present study, the highest value (4) was assigned to SPEC 1 species (major

concern) and the lowest (0) to Non-SPEC species (least concern). The SPEC score

was the sum of these values for each different species detected in each transect.

We built averaged mixed models to analyze the response variables. To

select the fixed factors in the model, we firstly performed a Principal Component

Analysis (PCA), with the Varimax Normalized factor rotation, with all plausible

variables in each season (Appendix S3) to explore the degree of association among

variables. The percents (%) of field types and vegetation cover were arcsine

square root transformed. We only considered axes with eigenvalue >2. We selected

the variable that correlated most strongly with the axis for further analyses

(always ≥0.7, Appendix S4) to reduce multicollinearity among variables

(Barrientos and Arroyo 2014). We preferred to use raw variables instead of PCA

factors because their meaning is easier to interpret (Barrientos 2010). This

technique allows highly correlated variables to be discarded (which can also be

done with simple correlations) and objectively selects the most biologically

meaningful and influential variable with each factor (Barrientos 2010). Secondly,

we followed the procedure described in the study of Zuur et al. (2009) to select the

random factor in mixed models. We built the most complex model (beyond optimal

model) with all fixed factors from PCA and including different random factors. We

considered year, site, both combined, or year nested within site as plausible

random factors. To select the random factor, we used the results from the ANOVA

132

test in R-program (R Core Team 2013). We built all possible models for the four

response variables in each season using a subsample of the data (154 cases for

each season, 70% of the dataset), leaving the rest of data for the validation process.

We selected models with an increase in corrected Akaike0s Information Criterion

(ΔAICc) over the best model <5 as candidate models (Burnham and Anderson

2002). Finally, with all these models, we performed an average model estimation,

with the package MumIN (Barton 2013) in Rprogram, in which the parameter

estimates of all models were combined (Burnham and Anderson 2002). The

random factor was that previously selected and the error structure was Poisson

for abundance, richness and SPEC-score and Gaussian for diversity. The final

averaged models included those variables identified as significant (those whose

confident interval excluded 0 value; Alonso et al. 2012; Delgado et al. 2013).

We developed two sets of models for each response variable. The first was

that of food models which used detailed seed and arthropod biomass

measurements and parameters describing the vegetation structure and the

surfaces of each field type as candidate variables (see Appendix S3). The second

was that of habitat models, built using field types and surfaces (Appendix S3). In all

cases, the variables selected from each axis had a correlation value ≥0.7 (Appendix

S4). We explored the predictive power of each set of models on the 30% previously

discarded data set (66 transects distributed evenly among all study sites in each

season). This validation shows how accurately the best model predicts data not

used before (Seiler 2005; Vaughan and Ormerod 2005). In spite of the

acknowledged importance of model validation in behavioral and ecological studies

in general, and distribution modeling studies in particular, this issue has been

generally ignored in the literature analyzing the efficiency of AES programs.

RESULTS

Wintering season

In winter food models, bird abundance was positively predicted by seed biomass

found in AES legume fields, surface of AES fallows and ploughed fields, and surface

of non-AES cereal stubbles (Table 2). The surface of AES fallows was the most

133

important variable. Bird richness was determined by AES fallow surface, total

arthropod biomass, and seed biomass from non-AES high-quality fields (stubbles,

fallows, and legume fields; Table 2). Again, the most important variable was the

surface of AES fallows. Bird diversity was best explained by a model including the

surface of ploughed and cereal stubble fields from regular farming activity (notice:

in this case the influence was negative), and the seed biomass from non-AES high-

quality fields (non-AES stubbles, fallows, and legumes; Table 2). The SPEC score

was determined by seed biomass in legume fields and surface of fallows, in both

cases from AES. The standardized regression coefficients showed that AES fallow

surface was the most important variable predicting SPEC-score.

Table 2. Model-averaged estimates of the food models for bird abundance, richness, diversity and

SPEC-score during the wintering, mating and post- fledging periods. The statistics given are: sum of

Akaike weights of the models in which the predictor was retained (Σ), parameter estimate of the

regression equation (b), standard deviation of the regression parameter (SE), lower and upper

confident limits of b, and standardized coefficients of predictors (β). Non-significant predictors are

not included. Factors are ordered by magnitude of the β coefficient.

Period Variable Parameter Σ b SE Lower

CI Upper

CI β

Wintering Abundance Intercept 3.69 0.24 3.22 4.16 0.08

FallowAES 1 4.36 0.46 3.44 5.28 0.18

SeedLegAES 1 1.66 0.63 0.40 2.92 0.10

CerStubbleNAT 1 0.88 0.17 0.54 1.22 0.01

Plough 0.88 0.43 0.21 0.01 0.85 0.01

Richness Intercept 1.27 0.14 0.99 1.55 0.14

FallowAES 0.87 0.99 0.31 0.37 1.61 0.25

SeedHQFNAT 1 0.54 0.18 0.18 0.90 0.08

ArthrTot 0.68 0.19 0.07 0.05 0.33 0.01

Diversity Intercept 0.26 0.05 0.15 0.37 0.00

Plough 1 1.03 0.08 0.87 1.20 0.27

CerStubbleNAT 1 -0.47 0.10 -0.68 -0.27 -0.16

SeedHQFNAT 1 0.28 0.09 0.10 0.46 0.08

SPEC-score Intercept 2.54 0.16 2.22 2.85 0.10

FallowAES 1 2.71 0.29 2.13 3.28 0.19

SeedLegAES 1 1.30 0.32 0.66 1.94 0.10

Mating

Abundance Intercept 2.22 0.08 2.06 2.38 0.01

134

Period Variable Parameter Σ b SE Lower

CI Upper

CI β

CerStubbleAES 1 2.64 0.49 1.66 3.62 0.11

LegAES 1 1.84 0.52 0.80 2.88 0.08

FallowNAT 1 1.36 0.45 0.47 2.25 0.05

Plough2 1 0.85 0.38 0.09 1.62 0.03

Richness Intercept 2.30 0.15 2.01 2.59 0.19

LegAES 1 1.69 0.12 1.44 1.94 0.12

FallowNAT 1 1.30 0.16 0.99 1.61 0.11

LegNAT 0.76 0.69 0.18 0.34 1.05 0.07

Diversity Intercept 0.35 0.02 0.30 0.40 0.04

FallowNAT 1 0.41 0.09 0.22 0.60 0.19

LegNAT 0.89 0.28 0.09 0.10 0.47 0.12

SPEC-score Intercept 2.49 0.20 2.08 2.90 0.08

FallowNAT 1 0.95 0.40 0.15 1.75 0.06

Plough2 0.94 0.70 0.34 0.00 1.37 0.04

Post-fledging

Abundance Intercept 3.20 0.39 2.42 3.98 0.04

ArthrFallowNAT 0.85 1.93 0.86 0.21 3.65 0.06

Plough 0.90 -1.12 0.37 -1.86 -0.38 -0.01

LegStubbleAES 0.75 0.74 0.35 0.04 1.44 0.01

Richness Intercept 1.07 0.21 0.65 1.49 0.13

FallowAES 1 1.26 0.18 0.91 1.62 0.13

ArthrFallowNAT 0.94 0.67 0.29 0.08 1.26 0.11

Plough2 1 0.60 0.22 0.16 1.03 0.07

Diversity Intercept 0.23 0.04 0.16 0.30 0.03

FallowAES 0.89 0.30 0.10 0.09 0.50 0.12

Plough2 0.88 0.31 0.10 0.12 0.50 0.12

ArthrFallowNAT 0.65 0.24 0.11 0.02 0.46 0.11

SPEC-score Intercept 1.38 0.15 1.09 1.68 0.06

Plough2 1 1.50 0.57 0.36 2.64 0.24

ArthrFallowNAT 1 0.74 0.32 0.11 1.37 0.07

FallowAES 0.65 0.55 0.15 0.25 0.85 0.02

The habitat models for bird abundance and richness included the surface of

AES legumes and fallows as predictors, while models for richness and diversity

included the surface of non-AES fallows (Table 3). Bird abundance and SPEC score

135

were predicted by the same variables. Also, non-AES cereal stubbles had the

lowest effect. The surface of non-AES fallows had a positive effect on bird richness

and diversity. Diversity was also positively affected by the surface of ploughed

fields (Table 3). Natural or managed fallows had the highest standardized

coefficient in all habitat models.

Table 3. Model-averaged estimates of the habitat models for bird abundance, richness, diversity

and SPEC-score during the wintering, mating and post- fledging periods. The statistics given are:

sum of Akaike weights of the models in which the predictor was retained (Σ), parameter estimate of

the regression equation (b), standard deviation of the regression parameter (SE), lower and upper

confident limits of b, and standardized coefficients of predictors (β). Non-significant predictors are

not included.

Period Variable Parameter Σ b SE Lower

CI Upper

CI β

Wintering Abundance Intercept 4.70 0.13 4.45 4.96 0.05

FallowAES 1 2.27 0.56 1.14 3.39 0.12

LegAES 1 1.67 0.25 1.18 2.16 0.04

CerStubbleNAT 0.87 0.70 0.06 0.58 0.82 3.84E-03

Richness Intercept 1.27 0.25 0.77 1.78 0.26

FallowAES 1 0.68 0.29 0.10 1.27 0.16

LegAES 1 0.63 0.23 0.17 1.10 0.12

FallowNAT 1 0.56 0.26 0.04 1.07 0.12

Diversity Intercept 0.21 0.08 0.04 0.38 0.06

FallowNAT 1 0.28 0.12 0.04 0.53 0.11

Plough 1 0.25 0.11 0.03 0.47 0.09

SPEC-score Intercept 2.47 0.17 2.13 2.81 0.10

LegAES 1 1.00 0.19 0.62 1.37 0.05

FallowAES 1 0.59 0.26 0.07 1.11 0.04

CerStubbleNAT 1 0.19 0.08 0.03 0.35 3.78E-03

Matinga

SPEC-score Intercept 1.80 0.18 1.44 2.16 0.05

LegNAT 1 1.40 0.50 0.40 2.41 0.11

Plough 1 -0.52 0.14 -0.81 -0.23 -0.01

FallowNAT 1 0.39 0.19 0.02 0.77 0.01

Post-fledging

Abundance Intercept 1.80 0.18 1.44 2.16 0.01

FallowNAT 1 1.73 0.52 0.70 2.77 0.03

136

Period Variable Parameter Σ b SE Lower

CI Upper

CI β

Richness Intercept 0.88 0.11 0.67 1.09 0.05

FallowAES 1 0.97 0.27 0.42 1.51 0.15

FallowNAT 1 0.62 0.29 0.04 1.20 0.10

Plough2 0.75 0.59 0.25 0.09 1.09 0.08

Diversity Intercept 0.22 0.04 0.14 0.29 0.03

FallowAES 1 0.29 0.10 0.09 0.50 0.12

Plough2 1 0.24 0.11 0.02 0.46 0.11

FallowNAT 1 0.20 0.09 0.03 0.37 0.07

SPEC-score Intercept 1.37 0.14 1.08 1.66 0.06

FallowNAT 1 0.82 0.26 0.30 1.34 0.06

Plough2 1 0.57 0.20 0.17 0.98 0.03

a Food and habitat models retained the same variables

Mating season

During the mating season, four significant variables were retained in food models

for bird abundance and three for richness, and two diversity and SPEC score (Table

2). The four variables influencing bird abundance had positive effects. Most

important were the surfaces of cereal stubbles and legume fields from AES,

followed by the surfaces of non-AES fallows and ploughed fields with sprouted

weeds.

The best food models for bird richness and diversity included non-AES

fallow and non-AES legume surfaces, both showing similar importance (Table 2).

The model for bird richness also included AES legume fields as the most influential

variable. The averaged model best explaining SPEC score included surfaces of non-

AES fallows and ploughed fields with weeds (Table 2). The influence of both

variables was positive, fallow surface showing the largest effect.

Habitat models best explaining bird abundance, richness, and diversity

during the mating season retained the same predictors as food models (Table 3).

The model averaging process showed that three variables influenced SPEC score:

surface of non-AES fallows and legumes with positive effects, and surface of

ploughed fields, with slightly negative effects.

137

Postfledging season

The biomass of arthropods in non-AES fallows was included in final food models

for abundance, richness, diversity and SPEC-score (Table 2). The surface of

ploughed fields with sprouted weeds was also included in the SPEC-score model

with a higher regression coefficient value. Also, the final model for bird abundance

included a negative effect of ploughed surface. The final models explaining bird

richness and diversity shared all predictors, namely arthropod biomass in non-AES

fallows, surface of AES fallows, and surface of ploughed fields with sprouted

weeds. In all cases, the amount of AES fallows was the most important variable

(Table 2).

Non-AES fallow surface was present in all final habitat models and ploughed

land with sprouted weeds was absent only in the model for bird abundance (Table

3). The surface of AES fallows was retained in final models for bird richness and

diversity (Table 3). All variables had a positive influence, and those derived from

AES showed the highest importance when they were included in the models.

Comparison of models and their cost

The sensitivity analysis showed that food models had a consistently higher

predictive ability than habitat models. The average increase in fit to the data was

13%, reaching a 20% in some variables (Table 4). Fit values were highest for bird

abundance and lowest for SPEC-score in both, food and habitat models, and in all

three seasons. SPEC score models showed the highest differences in predictive

ability between food and habitat models (18% on average). Differences between

food and habitat models were usually higher during the postfledging season than

during the wintering season.

138

Table 4. Sensitivity analyses for testing the predictive ability of the food and habitat models

Model predictive ability (%)

Season Dependent variable

Food Habitat Relative

increase 1

Wintering

Abundance 61.8 54.8 12.8

Richness 56.7 49.3 15.0

Diversity 50.3 44.9 12.0

SPEC-score 41.5 34.6 19.9

Mating

Abundance 61.4 -2 -

Richness 52.7 -2 -

Diversity 42.9 -2 -

SPEC-score 39.8 34.6 15.0

Post- fledging

Abundance 51.4 42.4 21.2

Richness 50.3 41.8 20.3

Diversity 43.5 36.3 19.8

SPEC-score 38.4 32.3 18.9

SPEC, Species of European Conservation Concern.

a Calculated as (food / habitat) x 100

b Food and habitat models retained the same variables

Obtaining food models required almost 2800 h, which was 20 times more

than those needed for habitat models (Table 5). Most of this extra time was due to

field work (86%). Also, food models needed more than 16,000 € mainly due to the

salaries (65%), a cost five times higher than that of habitat models.

139

Table 5. Comparison of costs (in €, work hours or number of people) incurred to measure variables

used in food and habitat models

Effort Food models Habitat models

People (n) 2 1

Field time (h) 2422 135

Laboratory time (h) 372 0

Total time (h) 2794 135

Total time (days) 276 45

Salaries (€) 10750 2250

Fuel (€) 1830 675

Food (€) 2440 450

Field material a (€) 1286 0

Laboratory material b (€) 105 0

Total cost (€) 16411 3375

a Small sampling equipment only

b Small laboratory equipment only. A binocular loupe and two optical microscopes were provided

by the research institute

DISCUSSION

Food versus habitat models

Habitat and food averaged models retained similar predictor variables, and the

main conclusion from both sets of models was that AES benefited steppe birds by

increasing the responses analyzed. However, food models were more effective in

explaining bird responses, going beyond the simple assessment that AES measures

were favorable. First, at least in winter and during the post-fledging season, they

had a higher predictive power than habitat models (respectively, 15% and 20%

higher). Second, they helped inferring important details about the ultimate causes

underlying bird responses in different periods of the annual cycle. For example,

while during winter habitat models included the surface of fallows, stubbles and

legume fields as main predictors, food models revealed the specific importance of

seed biomass for most of the response variables. Food models also highlighted the

importance of arthropod biomass in fallows during the post-fledging season, when

invertebrates are a major component of the diet of juveniles in most bird species.

140

These results show that in two of three seasons birds responded primarily to the

amount of food rather than to the surface of fields or the vegetation structure,

which was not included in any model. These results support our hypothesis that

increasing food resources leads to significantly higher birds numbers for those

periods in our study area.

An estimate of the seed biomass, either only in legume fields or altogether

in the three substrates considered of high quality (legume fields, stubbles and

fallows), was retained in the best winter food models for all response variables

investigated. Food models thus captured the important role that seeds play as a

source of energy and nitrogen for steppe birds during winter (Evans et al., 2011).

The amount of seeds in AES legume fields was particularly important for

abundance and SPEC-score models. The positive effect observed on averaged

model for the SPEC score means that by increasing the offer of legume fields in

winter, the AES program contributed to enhance not only the overall abundance of

birds maybe by attracting of local birds to food, but particularly that of species

with special conservation interest. This result contradicts the findings og Kleijn et

al. (2006), who suggested than endangered species rarely benefited from AEM.

Richness and diversity of the bird community responded to the total seed biomass

in all high-quality substrates (fallows, cereal stubbles, and legume fields) including

non AES fields.

Food models identified the biomass of arthropods as a further variable

influencing species richness during winter. That the richness of the steppe bird

community was affected by arthropod biomass in winter was a quite unexpected

result, as during this season most birds are typically granivorous. This result

highlights the importance that arthropods may have even in winter for the bird

community of dry cereal farmland in Mediterranean latitudes, whose climatic

conditions may favor the presence of arthropod reservoirs available for wintering

bird species.

In sum, with the exception of the mating season food models were better

than habitat models in predicting bird community responses. In the latter, the

direct importance of seeds and arthropods would have gone unnoticed. The gain in

predictive power was highest for bird abundance and richness models in summer

141

(respectively, 21.2% and 20.3%), and the SPEC-score model in winter (19.9%).

The advantage of a higher predictive power of food models should, however, be

balanced against their much higher cost. In this study, the cost of obtaining data for

food models was five times higher than for habitat models. Two persons, 130 days

field work and 146 days laboratory work (ca. 3000 working hours in total) were

necessary to measure the biomass of arthropods and seeds, and the vegetation

structure variables. In addition, fuel, materials, and maintenance costs of personnel

were also higher in food models. A five times higher cost could appear to be an

excessive expenditure, but the additional cost of quantifying food availability may

only represent a minor fraction of the total cost of agri-environment programmes.

Our study calls attention to the fact that bird responses could remain

unexplained if they are judged only from an assessment of the habitat variables.

This could lead to erroneous AES efficiency assessments. In contrast,

measurements of food availability and vegetation structure could add important

details to help interpreting the reactions of the bird community to AES

interventions and thus facilitate evaluating the real efficiency of AES programs.

The decision whether to invest in such detailed measurements should be taken

considering the specific circumstances of each particular AES program.

Benefits of AES measures in different seasons

Considering all seasons together, the most effective AES measure was probably the

provision and maintenance of fallows. Fallow fields were among the most

significant predictors in a majority of food and habitat models in the three seasons.

Previous studies already highlighted the importance of fallows in providing food

and refuge for several steppe bird species (Duelli and Obrist 2003; Suárez et al.

2004; Billeter et al. 2008). Fallows are perhaps the only substrate offering

sufficient amount of varied food types including weeds, seeds, and arthropods

(Campbell et al. 1997; Herkert 2009; Lapiedra et al. 2011). It is therefore not

surprising that these substrates appear in many AES studies as critical to increase

bird abundance, richness and diversity.

Maintenance of cereal stubbles through the winter did not appear to be an

AES measure providing a significant benefit to steppe birds, probably because in

142

nonintensive dry cereal areas non-AES cereal stubbles are already abundant in

winter. A previous study (Suárez et al. 2004) also suggested that in Spain stubble

maintenance through the winter did not benefit farmland birds since these can

feed on various non-cultivated substrates. However, in areas where farming is

more intensive and thus cereal stubbles scarce or are usually absent in winter, an

AES including cereal stubble maintenance may certainly benefit steppebirds

(Gillings et al. 2005; Concepción and Díaz 2011; Concepción et al. 2012). In winter

food models, the surface of natural cereal stubbles had a positive effect on bird

abundance, but a negative effect on bird diversity. This could be so due to the

differences in diet among species (Princé et al. 2012), or simply because a marked

preponderance of a single species as skylark (Alauda arvensis, Appendix S1)

implies a reduction of species diversity values. Anyway, raising habitat quality by

increasing the amount of food in winter may also produce delayed benefits during

the following breeding season (Gillings et al. 2005). Several studies showed that

breeding success or fitness of some species were correlated with conditions

experienced during the preceding winter (Peach et al. 1999, 2001; Siriwardena et

al. 1999, 2000, 2007).

The positive effect of ploughed fields on winter bird abundance and

diversity was surprising, given the low values for vegetation cover and arthropod

and seed biomass typical of these substrates (this study; see also Díaz and Tellería

1994). However, other authors qualified ploughed fields as important for some

bird species (Suárez et al. 2004). For example, wagtails (Motacilla spp.) or cattle

egrets (Bubulcus ibis) follow tractors to feed on invertebrates unearthed during

ploughing. A similar behavior has been described for certain granivorous birds

that take unearthed seeds (e.g., Whittingham et al. 2006). Finally, certain species

may be favored by the lower vegetation cover and the consequent higher

antipredator visibility in ploughed fields (Butler et al. 2005), although this

possibility has not been tested in this study.

During the mating season, habitat and food models basically coincided. The

retained variables were surfaces of various substrates, but no food biomass

estimates were included as predictors. This result contrasts with some previous

studies (Traba et al. 2008; Concepción and Díaz 2010) which suggest that food

143

availability is a key factor during the mating season. We can think of two possible

reasons for the absence of a significant effect of biomass variables during this

period of the annual cycle. First, during the mating season, weeds and

invertebrates are more abundant in our non-intensive farmland area compared

with other periods of the year. Food availability may then be higher than demands

and thus represent no limiting factor for bird abundance and diversity.

Consequently, the effect of food biomass in each particular substrate type may be

obscured during this season. Second, during the mating season, most birds are

involved in defending territories, pairing and searching for nest sites. For these

tasks, a rough estimate of available surfaces of the different field types may be a

better indicator of habitat suitability than a precise estimate of food availability.

Substrate selection during this season may indeed provide enough information

about the best place to nest and the food availability for chicks expected later in

the season. Third, bird abundance, diversity, and richness in spring are limited by

other variables such as territoriality, and complex intra and interspecific

interactions within the breeding bird community. In spite of the absence of a clear

effect of food biomass during the mating season, the surface of legumes was

correlated with the biomass of arthropods in this substrate (0.81), showing its role

in increasing food availability. Previous studies had already described the

importance of legumes as a source of nitrogen for many species (Karasov 1990),

and in particular for steppe birds in dry farmland areas (Bretagnolle et al. 2011;

Bravo et al. 2012). Our study suggests that legumes may also be important as a

food source for insectivorous species.

Unlike during winter, AES cereal stubbles were the most important

predictor in spring averaged model for bird abundance. For birds, natural and AES-

managed cereal stubbles are probably identical during most part of the winter, but

on managed stubbles, AES restrictions prevent herbicide and insecticide use

between harvest and ploughing. This fact surely favored the growth of abundant

weeds on managed stubbles in spring and released the observed positive response

from birds to the AES stubbles during the mating season. According to the AES

rules, ploughing is allowed in January–March, but later forbidden until July. This

was an unexpected positive effect of the AES, as we thought all farmers would

plough AES stubbles before the deadline of 31 March, as they did with non-AES

144

stubbles (which are also sprayed usually with chemical products against weeds). It

is necessary to highlight that AES stubbles were kept from spring to the following

autumn without any additional payments to farmers.

During the postfledging season, habitat and food models were also different.

While in habitat models, the surface of fallows and ploughed fields with abundant

sprouted weeds were the main predictors, food models showed that the birds’

response was really induced by the biomass of arthropods in non-AES fallows. In

food models, the surface of ploughed fields with sprouted weeds is still retained as

a significant predictor. This is because in dry cereal farmland areas ploughed fields

with sprouted weeds are the only substrate where birds can find green plants and

associated canopy arthropods in summer.

CONCLUSIONS

Our study showed that the AES contributed to increase the abundance and

diversity of farmland birds in our study area. The positive responses observed in

four variables analyzed were in part induced by some of the AES measures applied.

However, many land-use variables not regulated by the AES were also important,

probably due to the extensive agricultural regime predominant in our study area.

The models presented in our study enabled evaluating the percent benefit

obtained from AES measures as compared to non-AES land-use variables.

Exhaustive field work was devoted in this study to measure landscape

complexity, vegetation structure and food availability, all of which are considered

important factors influencing bird behavior and distribution patterns. In our study,

various important effects of seed and arthropod biomass detected using food

models would have gone unnoticed using habitat models where only substrate

composition is measured. This was at the cost of a much higher investment in time

and personnel in food models, with a consequent increase in the total cost of the

research. However, detailed food measurements allowed increasing the

explanatory power of models describing bird responses, as well as identifying the

causes underlying these responses.

145

Our study finally highlighted the need to apply AES measures and to study

bird responses separately in different periods of the year. As most birds use

different substrates throughout the annual cycle, the same AES measures may have

different effects in different seasons. Thus, farmland areas need to be managed

from that seasonal perspective to maximize the benefits of AES programs. In our

case, the AES measures aimed at enhancing the benefits of traditional farming

cycles at dry cereal areas, by providing supplementary legume crops and fallows

and limiting tillage work.

ACKNOWLEDGEMENTS

We thank L. M. Bautista, J. M. Álvarez, R. Early and R. Barrientos for their help

during statistical analyses and comments on the manuscript. We thank farmers in

the study area for their collaboration. M. Magaña was the CSIC manager of the AES

during part of the study. Several students from the Universidad Autónoma de

Madrid and Universidad Complutense de Madrid helped during field work. We

especially thank Alberto Lucas who intensively collaborated during arthropod and

seed sampling. Compensatory payments to farmers were financed through a

contract CSIC-HENARSA. The study was financed by the Dirección General de

Investigación of the Spanish Ministry for Science and Innovation (project

CGL2008-02567, with contributions from project CGL2005-04893).

REFERENCES

Alonso, J. C., J. M. Álvarez, and C. Palacín. 2012. Leks in ground-displaying birds:

hotspots or safe places? Behav. Ecol. 23:491–501.

Barrientos, R. 2010. Retention of native vegetation within the plantation matrix

improves its conservation value for a generalist woodpecker. Forest Ecol.

Manag. 260:595–602.

Barrientos, R., and B. Arroyo. 2014. Nesting habitat selection of Mediterranean

raptors in managed pinewoods: searching for common patterns to derive

conservation recommendations. Bird Conserv. Int. 24:138–151.

146

Barton, K., 2013. MuMIn: Multi-model inference. R package version 1.9.13.

Benton, T. G., J. A. Vickery, and J. D. Wilson. 2003. Farmland biodiversity: is habitat

heterogeneity the key? Trends Ecol. Evol. 18:182–188.

Billeter, R., J. Liira, D. Bailey, R. Bugter, P. Arens, I. Augenstein, et al. 2008.

Indicators for biodiversity in agricultural landscapes: a pan-European

study. J. Appl. Ecol. 45:141–150.

BirdLife International. 2004. Birds in Europe: population estimates, trends and

conservation status. BirdLife International, Cambridge, U.K.

Boyce, M. S., and L. L. McDonald. 1999. Relating populations to habitats using

resource selection functions. Trends Ecol. Evol. 14:268–272.

Bravo, C., C. Ponce, C. Palacín, and J. C. Alonso. 2012. Diet of young Great Bustards

Otis tarda in Spain: sexual and seasonal differences. Bird Study 59:243–251.

Breeuwer, A., F. Berendse, F. Willems, R. Foppen, W. Teunissen, H. Schekkerman, et

al. 2009. Do meadow birds profit from agri-environment schemes in Dutch

agricultural landscapes? Biol. Conserv. 142:2949–2953.

Bretagnolle, V., A. Villers, L. Denonfoux, T. Cornulier, P. Inchausti, and I.

Badenhausser. 2011. Rapid recovery of a depleted population of Little

Bustards Tetrax tetrax following provision of alfalfa through an agri-

environment scheme. Ibis 153:4–13.

Burnham, K. P., and D. R. Anderson. 2002. Model selection and multi-model

inference. A practical information-theoretic approach. Springer, New York.

Butler, S. J., and K. Norris. 2013. Functional space and the population

dynamics of birds in agro-ecosystems. Agr. Ecosyst. Environ., 164:200–208.

Butler, S. J., R. B. Bradbury, and M. J. Whittingham. 2005. Stubble height affects the

use of stubble fields by farmland birds. J. Appl. Ecol. 42:469–476.

Butler, S. J., J. A. Vickery, and K. Norris. 2007. Farmland biodiversity and the

footprint of agriculture. Science 315:381–384.

147

Butler, S. J., L. Boccaccio, R. D. Gregory, P. Vorisek, and K. Norris. 2010a. Quantifying

the impact of land-use change to European farmland bird populations. Agr.

Ecosyst. Environ., 137:348–357.

Butler, S. J., E. H. A. Mattison, N. J. Glithero, L. J. Robinson, P. W. Atkison, S. Gillings,

et al. 2010b. Resource availability and the persistence of seed-eating bird

populations in agricultural landscapes: a mechanistic modelling approach. J.

Appl. Ecol. 47:67–75.

Campbell, L. H., M. J. Avery, P. Donald, A. D. Evans, R. E. Green, and J. D. Wilson.

1997. A review of the indirect effects of pesticides on birds. JNCC Report No.

227. JNCC, Peterborough.

Cardinale, B. J., J. E. Duffy, A. Gonzalez, D. U. Hooper, C. Perrings, P. Venail, et al.

2012. Biodiversity loss and its impact on humanity. Nature 486:59–67.

Chamberlain, D. E., R. J. Fuller, J. C. Bunce, J. C. Duckworth, and M. Shrubb. 2000.

Changes in the abundance of farmland birds in relation to the timing of

agricultural intensification in England and Wales. J. Appl. Ecol. 37:771–788.

Concepción, E. D., and M. Díaz. 2010. Relative effects of field and landscape-scale

intensification on farmland bird diversity in Mediterranean dry cereal

croplands. Asp. Appl. Biol. 100:245–252.

Concepción, E. D., and M. Díaz. 2011. Field, landscape and regional effects of

farmland management on specialist open-land birds: does body size

matter? Agr. Ecosyst. Environ. 142:303–310.

Concepción, E. D., M. Díaz, and R. A. Baquero. 2008. Effects of landscape complexity

on the ecological effectiveness of agri-environment schemes. Landscape

Ecol. 23:135–148.

Concepción, E. D., M. Díaz, D. Kleijn, A. Báldi, P. Batáry, Y. Clough, et al. 2012.

Interactive effects of landscape context constrain the effectiveness of local

agri-environmental management. J. Appl. Ecol. 49:695–705.

148

Davey, C. M., J. A. Vickery, N. D. Boatman, D. E. Chamberlain, H. R. Parry, and G. M.

Siriwardena. 2010. Assessing the impact of entry level stewardship on

lowland farmland birds in England. Ibis 152:459–474.

Delgado, M. P., M. Sanza, M. B. Morales, J. Traba, and D. Rivera. 2013. Habitat

selection and coexistence in wintering passerine steppe birds. J. Ornithol.

154:469–479.

Díaz, M., and J. L. Tellería. 1994. Predicting the effects of agricultural changes in

central Spain croplands on seed-eating overwintering birds. Agr. Ecosyst.

Environ. 49:289–298.

Díaz, M., E. D. Concepción, and R. A. Baquero. 2006. Direct and indirect methods for

the design and evaluation of European agri-environment schemes. J.

Ornithol. 147:156–157.

Donald, P. F., R. E. Green, and M. F. Heath. 2001. Agricultural intensification and the

collapse of Europe’s farmland bird populations. Proc. R. Soc. Lond. B

268:25–29.

Duelli, P., and M. K. Obrist. 2003. Regional biodiversity in an agricultural landscape:

the contribution of semi-natural habitat islands. Basic Appl. Ecol. 4:129–

138.

EBCC, 2010. Population Trends of European Common Birds 2010. EBCC report.

http://www.ebcc.info/index.php?ID=392 (accessed 26 June 2013).

Ekroos, J., O. Olsson, M. Rundlo€f, F. W€atzold, and H. G. Smith. 2014. Optimizing

agri-environment schemes for biodiversity, ecosystem services or both?

Biol. Conserv. 172:65–71.

Escandell, V., D. Palomino, B. Molina, A. Leal, C. Remacha, A. Bermejo, et al. 2011.

Programas de seguimiento de SEO/ BirdLife en 2009–2010. SEO/BirdLife,

Madrid.

Evans, D. M., J. O. Pocock, J. Brooks, and J. Memmott. 2011. Seeds in farmland food-

webs: resource importance, distribution and the impacts of farm

management. Biol. Conserv. 144:2941–2950.

149

Gabriel, D., S. M. Sait, J. A. Hodgson, U. Schmutz, W. E. Kunin, and T. G. Benton. 2010.

Scale matters: the impact of organic farming on biodiversity at different

spatial scales. Ecol. Lett. 13:858–869.

Gillings, S., S. E. Newson, D. G. Noble, and J. A. Vickery. 2005. Winter availability of

cereal stubbles attracts declining farmland birds and positively influences

breeding population trends. Proc. Biol. Sci. 272:733–739.

Hanspach, J., J. Fischer, J. Stott, and K. Stagoll. 2011. Conservation management of

eastern Australian farmland birds in relation to landscape gradients. J. Appl.

Ecol. 48:523–531.

Henderson, I. G., J. Cooper, R. J. Fuller, and J. Vickery. 2000. The relative abundance

of birds on set-aside and neighbouring fields in summer. J. Appl. Ecol.

37:335–347.

Herkert, J. R. 2009. Response of bird populations to farmland set-aside programs.

Conserv. Biol. 23:1036–1040.

Hódar, J. A. 1996. The use of regression equations for estimation of arthropod

biomass in ecological studies. Acta Oecol. 17:421–433.

Hooper, D. U., E. C. Adair, J. Cardinale, J. E. K. Byrnes, B. A. Hungate, K. L. Matulich, et

al. 2012. A global synthesis reveals biodiversity loss as a major driver of

ecosystem change. Nature 486:105–108.

Karasov, W. H. 1990. Digestion in birds: chemical and physiological determinants

and ecological implications. Pp. 391–415 in M. L. Morrison, C. J. Ralph, J.

Verner and J. R. Jehl, eds. Avian foraging: theory, methodology and

applications. Cooper Ornithological Society, San Diego, CA.

Kleijn, D., and W. J. Sutherland. 2003. How effective are European agri-

environment schemes in conserving and promoting biodiversity? J. Appl.

Ecol. 40:947–969.

Kleijn, D., R. A. Baquero, Y. Clough, et al. 2006. Mixed biodiversity benefits of agri-

environment schemes in five European countries. Ecol. Lett. 9:243–254.

150

Kleijn, D., M. Rundlöf, J. Scheper, H. G. Smith, and T. Tscharntke. 2011. Does

conservation on farmland contribute to halting the biodiversity decline?

Trends Ecol. Evol. 26:474–481.

Lane, S. J., J. C. Alonso, J. A. Alonso, and M. A. Naveso. 1999. Seasonal changes in diet

and diet selection of great bustards (Otis t. tarda) in north-west Spain. J.

Zool. Lond. 247:201–214.

Lapiedra, O., A. Ponjoan, A. Gamero, G. Bota, and S. Mañosa. 2011. Brood ranging

behaviour and breeding success of the threatened little bustard in an

intensified cereal farmland area. Biol. Conserv. 144:2882–2890.

MacDonald, M. A., M. Maniakowski, G. Cobbold, P. V. Grice, and G. Q. Anderson.

2012. Effects of agri-environment management for stone curlews on other

biodiversity. Biol. Conserv. 148:134–145.

Marfil-Daza, C., M. Pizarro, and G. Moreno-Rueda. 2013. Do hot spots of breeding

birds serve as surrogate hot spots of wintering birds? An example from

central Spain. Anim. Conserv. 16:60–68.

Miyake, S., M. Renouf, A. Peterson, C. McAlpine, and C. Smith. 2012. Land-use and

environmental pressures resulting from current and future bioenergy crop

expansion: a review. J. Rural Stud. 28:650–658.

Newton, I. 2004. The recent declines of farmland bird populations in Britain: an

appraisal of causal factors and conservation actions. Ibis 146:579–600.

Oliver, I., and A. J. Beattie. 1993. A possible method for the rapid assessment of

biodiversity. Conserv. Biol. 7:562–568.

Peach, W. J., G. M. Siriwardena, and R. Gregory. 1999. Long-term changes in over-

winter survival rates explain the decline in Reed Buntings Emberiza

Schoeniclus in Britain. J. Appl. Ecol. 36:798–811.

Peach, W. J., L. J. Lovett, S. R. Wotton, and C. Jeffs. 2001. Countryside stewardship

delivers cirl bunting in Devon. Biol. Conserv. 101:361–373.

151

Perkins, A. J., H. E. Maggs, A. Watson, and J. D. Wilson. 2011. Adaptive management

and targeting of agri-environment schemes does benefit biodiversity: a case

study of the corn bunting Emberiza calandra. J. Appl. Ecol. 48:514–522.

Ponce, C., C. Bravo, D. García de León, M. Magaña, and J. C. Alonso. 2011. Effects of

organic farming on plant and arthropod communities: a case study in

Mediterranean dryland cereal. Agr. Ecosyst. Environ. 141:193–201.

Princé, K., J. P. Moussus, and F. Jiguet. 2012. Mixed effectiveness of French agri-

environment schemes for nationwide farmland bird conservation. Agr.

Ecosyst. Environ. 149:74–79.

R Core Team, 2013. R: A language and environment for statistical computing. R

Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0,

URL http://www.R-project.org/.

Robinson, R. A., and W. J. Sutherland. 2002. Post war changes in arable farming and

biodiversity in Great Britain. J. Appl. Ecol. 39:157–176.

Seiler, A. 2005. Predicting locations of moose-vehicle collisions in Sweden. J. Appl.

Ecol. 42:371–382.

Siriwardena, G. M., S. R. Baillie, and J. D. Wilson. 1999. Temporal variation in the

annual survival rates of six granivorous birds with contrasting population

trends. Ibis 141:621–636.

Siriwardena, G. M., S. R. Baillie, H. Q. P. Crick, and J. D. Wilson. 2000. The

importance of variation in the breeding performance of seed-eating birds in

determining their population trends on farmland. J. Appl. Ecol. 37:128–148.

Siriwardena, G. M., D. K. Stevens, G. Q. A. Anderson, J. A. Vickery, N. A. Calbrade, and

S. Dodd. 2007. The effect of supplementary winter seed food on breeding

populations of farmland birds: evidence from two large-scale experiments.

J. Appl. Ecol. 44:920–932.

Stoate, C., I. G. Henderson, and D. M. B. Parish. 2004. Development of an agri-

environment scheme option: seed-bearing crops for farmland birds. Ibis

146:203–209.

152

Stoate, C., A. Báldi, P. Beja, N. D. Boatman, I. Herzon, A. van Doorn, et al. 2009.

Ecological impact of early 21st century agricultural change in Europe—a

review. J. Environ. Manage. 91:22–46.

Suárez, F., V. Garza, J. J. Oñate, E. L. García de la Morena, A. Ramírez, and M. B.

Morales. 2004. Adequacy of winter stubble maintenance for steppe

passerine conservation in central Spain. Agr. Ecosyst. Environ. 104:667–

671.

Tilman, D., C. Balzer, J. Hill, and B. L. Befort. 2011. Global food demand and the

sustainable intensification of agriculture. Proc. Natl Acad. Sci. U S A

108:20260–20264.

Tilman, D., P. Reich, and F. Isbell. 2012. Biodiversity impacts ecosystem

productivity as much as resources, disturbance or herbivory. Proc. Natl

Acad. Sci. U S A 109: 10394–10397.

Traba, J., M. B. Morales, E. L. García de la Morena, and M. P. Delgado. 2008. Selection

of breeding territory by little bustard (Tetrax tetrax) males in Central Spain:

the role of arthropod availability. Ecol. Res. 23:615–622.

Tscharntke, T., A. M. Klein, A. Kruess, I. Steffan-Dewenter, and C. Thies. 2005.

Landscape perspectives on agricultural intensification and biodiversity-

ecosystem service management. Ecol. Lett. 8:857–874.

Vaughan, I. P., and S. J. Ormerod. 2005. The continuing challenges of testing species

distribution models. J. Appl. Ecol. 42:720–730.

Verhulst, J., D. Kleijn, and F. Berendse. 2007. Direct and indirect effects of the most

widely implemented Dutch agri-environment schemes on breeding waders.

J. Appl. Ecol. 44:70–80.

Vickery, J. A., J. R. Tallowin, R. E. Feber, E. J. Asteraki, P. W. Atkinson, R. J. Fuller, et

al. 2001. The management of lowland neutral grasslands in Britain: effects

of agricultural practices on birds and their food resources. J. Appl. Ecol.

38:647–664.

153

Whittingham, M. J., C. L. Devereux, A. D. Evans, and R. B. Bradbury. 2006. Altering

perceived predation risk and food availability: management prescriptions

to benefit farmland birds on stubble fields. J. Appl. Ecol. 43:640–650.

Wilson, J. D., A. D. Evans, and P. V. Grice. 2009. Bird conservation and agriculture.

Cambridge Univ. Press, Cambridge, U.K.

Winqvist, C., J. Bengtsson, T. Aavik, F. Berendse, et al. 2011. Mixed effects of organic

farming and landscape complexity on farmland biodiversity and biological

control potential across Europe. J. Appl. Ecol. 48:570–579.

Wuczyński, A., K. Kujawa, Z. Dajdok, and W. Grzesiak. 2011. Species richness and

composition of bird communities in various field margins of Poland. Agr.

Ecosyst. Environ. 141:202–209.

Zuur, A. F., E. N. Ieno, N. J. Walker, A. A. Saveliev, and G. M. Smith. 2009. Mixed

effects models and extensions in ecology with R. Springer, New York.

154

Appendix S1. List of bird species (in alphabetical order) contacted during surveys of the study area,

indicating their total abundance (sum of abundance values in all transects) in each season and the

Species of European Conservation Concern (SPEC) category.

Species SPEC Wintering Mating Post-fledging

Accipiter gentilis NON-SPEC

1

Alauda arvensis SPEC 3 8625 6 5

Alectoris rufa SPEC 2 185 108 96

Anas platyrhynchos NON-SPEC

21 1

Anthus campestris SPEC 3

14

Anthus pratensis NON-SPEC 246

Apus apus NON-SPEC

2 9

Aquila adalberti SPEC 1

1

Asio flammeus SPEC 3 7

Athene noctua SPEC 3

3

Bubo bubo SPEC 3

1

Bubulcus ibis NON-SPEC 29 16 185

Burhinus oedicnemus SPEC 3

66 110

Buteo buteo NON-SPEC 10 3 5

Calandrella brachydactila SPEC 3

72

Carduelis cannabina SPEC 2 967 110 77

Carduelis carduelis NON-SPEC 362 75 121

Carduelis chloris NON-SPEC 99 43 11

Ciconia ciconia SPEC 2

4 16

Circaetus gallicus SPEC 3

5

Circus aeruginosus NON-SPEC 7 9 6

Circus cyaneus SPEC 3 8 6 3

Circus pygargus NON-SPEC

21 19

Cisticola juncidis NON-SPEC 5 74 52

Clamator glandarius NON-SPEC 3 17 2

Columba livia NON-SPEC 120 98 461

Columba palumbus NON-SPEC 33 84 395

Corvus corax NON-SPEC

3 1

Corvus corone NON-SPEC 4

7

Corvus monedula NON-SPEC 29 25 14

Coturnix coturnix SPEC 3 3 102 16

Delichon urbicum SPEC 3

46 45

Emberiza schoeniclus NON-SPEC 4

Falco columbarius NON-SPEC 7

Falco naumanni SPEC 1

11

Falco peregrinus NON-SPEC

1

Falco subbuteo NON-SPEC

3

Falco tinnunculus SPEC 3 10 9 24

Fringilla coelebs NON-SPEC 84 6

Galerida cristata SPEC 3 139 265 336

Gelochelidon nilotica SPEC 3

10

Hieraaetus pennatus SPEC 3

1

Hippolais polyglotta NON-SPEC

1

Hirundo daurica NON-SPEC

1

Hirundo rustica SPEC 3

81 43

Lanius meridionalis SPEC 3 12 4 12

Lanius senator SPEC 2

3 16

Luscinia megarhynchos NON-SPEC

1

155

Species SPEC Wintering Mating Post-fledging

Melanocorypha calandra SPEC 3 524 660 712

Merops apiaster SPEC 3

7 16

Miliaria calandra SPEC 2 648 564 117

Milvus migrans SPEC 3

5 14

Milvus milvus SPEC 2 4

Motacilla alba NON-SPEC 98

Motacilla cinerea NON-SPEC

11

Motacilla flava NON-SPEC

30

Oenanthe hispanica SPEC 2

78 64

Oenanthe oenanthe SPEC 3

48 31

Otis tarda SPEC 1 541 280 160

Parus caeruleus NON-SPEC 1

Parus major NON-SPEC

1

Passer domesticus SPEC 3 305 118 165

Passer hispaniolensis NON-SPEC 80 52 139

Passer montanus SPEC 3 142 82 61

Petronia petronia NON-SPEC 56 42 43

Phoenicurus ochruros NON-SPEC 12

Phoenicurus phoenicurus SPEC 2

1

Phylloscopus collybita NON-SPEC 4 1

Pica pica NON-SPEC 169 192 241

Picus viridis SPEC 2 1

1

Pluvialis apricaria NON-SPEC 41

Pterocles alchata SPEC 3 12 2 9

Pterocles orientalis SPEC 3 84 28 7

Saxicola rubetra NON-SPEC

1

Saxicola torquata NON-SPEC 6 2 14

Serinus serinus NON-SPEC 110 98 95

Sturnus unicolor NON-SPEC 740 186 1086

Sylvia cantillans NON-SPEC

1 3

Sylvia melanocephala NON-SPEC 6 3 3

Tetrax tetrax SPEC 1 225 140 63

Tringa ochropus NON-SPEC

1

Turdus merula NON-SPEC 3

Turdus philomelos NON-SPEC 31

Upupa epops SPEC 3 1 9 16

Vanellus vanellus SPEC 3 368 15 1

Total birds 15210 4066 5174

156

Appendix S2. Correlation values between consecutive transects for each season. Autocorrelation coefficient values (ρ) calculated between each pair of

consecutive transects (lag=1) are shown. Asterisks show 0.05>P-values>0.01. There was no autocorrelation coefficient with a P-value <0.01. In each season

64 autocorrelation coefficients were calculated. Because this is a large number of tests, it was expected that a few of them would be significant by chance.

Therefore, we applied the Bonferroni correction to the P-values ( Bonferroni-corrected P values were 0.05/64=0.00078) and observed no significant

autocorrelation values after correction.

1) Wintering season. Four (6.3%) out of 64 coefficients (marked with *) were significant before Bonferroni correction

Abundance Richness Diversity Spec-score

Site 2006 2007 2008 2009 2006 2007 2008 2009 2006 2007 2008 2009 2006 2007 2008 2009

1 0.16 -0.52 -0.55 0.26 0.60* -0.07 -0.56 0.31 0.15 -0.10 -0.16 0.30 0.55 -0.70* 0.03 0.21

2 0.02 0.43 -0.07 0.05 0.01 -0.06 -0.19 0.19 0.03 0.05 -0.21 -0.21 -0.05 0.11 -0.02 0.05

3 0.13 -0.51 -0.26 -0.27 0.64* -0.06 0.23 -0.06 -0.20 0.20 0.31 0.20 0.26 -0.78* 0.15 -0.30

4 -0.36 0.10 -0.04 -0.05 0.18 -0.03 -0.15 -0.03 -0.19 -0.16 0.21 0.23 -0.35 0.09 -0.05 -0.06

2) Mating season. Two (3.1%) out of 64 coefficients (marked with *) were significant before Bonferroni correction

Abundance Richness Diversity SPEC-score

Site 2006 2007 2008 2009 2006 2007 2008 2009 2006 2007 2008 2009 2006 2007 2008 2009

1 -0.42 0.35 0.24 0.01 -0.33 0.63* 0.00 -0.05 -0.22 0.24 -0.11 -0.08 -0.02 0.51 0.34 0.48

2 0.42 0.10 -0.85* -0.08 0.31 -0.51 -0.02 -0.18 0.31 -0.59 0.20 -0.11 -0.30 -0.58 0.36 0.25

3 -0.50 0.07 -0.57 -0.33 -0.66 0.41 -0.02 0.06 -0.50 0.25 0.05 0.43 -0.17 0.03 -0.45 -0.32

4 -0.50 -0.37 -0.28 -0.06 0.00 0.18 0.13 -0.31 0.10 0.22 0.15 -0.31 -0.50 -0.37 -0.28 -0.06

157

3) Post-fledging season. One (1.6%) out of 64 coefficients (marked with *) were significant before Bonferroni correction

Abundance Richness Diversity SPEC-score

Site 2006 2007 2008 2009 2006 2007 2008 2009 2006 2007 2008 2009 2006 2007 2008 2009

1 0.14 -0.03 0.63* -0.53 -0.29 0.59 0.15 0.18 -0.38 -0.23 0.15 -0.23 -0.48 0.41 -0.24 -0.53

2 -0.19 -0.44 -0.14 -0.10 -0.15 -0.13 -0.22 0.22 0.14 0.37 0.06 0.33 0.13 0.08 -0.09 -0.10

3 -0.12 -0.17 -0.21 -0.40 -0.29 0.22 -0.18 0.22 -0.22 -0.13 -0.01 -0.13 0.03 0.01 -0.29 -0.40

4 -0.57 -0.26 0.35 -0.23 -0.35 -0.23 0.40 0.33 -0.23 -0.37 0.52 0.48 -0.52 -0.26 0.23 -0.23

158

Appendix S3. Variables defined to measure habitat characteristics and biomass of arthropods and

seeds, indicating the model type (food o habitat) in which they were included. Surface variables

were measured as percentages per transect, biomass variables as grams per transect, and height as

centimetres per transect. All vegetation and ground structure variables calculated for each transect

were derived from the mean value in each field type and site and according to its surface.

Short name Definition Model

ArthrAES Total arthropod biomass in the sum of all AES fallows,+ AES cereal stubbles + AES legume fields + AES legume stubble fields

Food

ArthrCerNAT Arthropod biomass in non-AES cereal fields Food

ArthrCerStubbleAES Arthropod biomass in AES cereal stubble fields Food

ArthrCerStubbleNAT Arthropod biomass in non-AES cereal stubble fields Food

ArthrEdge Arthropod biomass in edges Food

AES Surface of AES Habitat

ArthrFallowAES Arthropod biomass in AES fallow fields Food

ArthrFallowNAT Arthropod biomass in non-AES fallow fields Food

ArthrHQFNAT Arthropod biomass in the sum of all non-AES ‘high quality fields’ (non-AES fallows + non-AES cereal stubble + non-AES legumes + non-AES legume stubbles)

Food

ArthrLegAES Arthropod biomass in AES legume fields Food

ArthrLegNAT Arthropod biomass in non-AES legume fields Food

ArthrLegStubbleAES Arthropod biomass in AES legume stubble fields Food

ArthrLegStubbleNAT Arthropod biomass in non-AES legume stubble fields Food

ArthrNAT Arthropod biomass in non-AES fields Food

ArthrPlough Arthropod biomass in ploughed fields Food

ArthrPlough2 Arthropod biomass in ploughed fields with sprouted weeds

Food

ArthrTot Total arthropod biomass Food

CerNAT Surface of cereal fields Habitat

CerStubbleAES Surface of cereal stubble fields included in AES Habitat

CerStubbleNAT Surface of cereal stubble fields not included in AES Habitat

CerStubbleTot Total surface of cereal stubble Habitat

CoverAES Mean vegetation cover derived from AES in percentage Food

CoverTOTAL Mean vegetation cover Food

DifHeight Difference between the maximum and minimum vegetation height in cm

Food

Edge Surface of edges Habitat

FallowAES Surface of fallow fields included in AES Habitat

FallowNAT Surface of fallow fields not included in AES Habitat

FallowTot Total surface of fallow fields Habitat

HQFNAT Surface of non-AES ‘high quality fields’ (non-AES fallows + non-AES cereal stubble + non-AES legumes + non-AES legume stubbles)

Habitat

LandscapeDiversity Landscape diversity (Shannon index) Habitat

LandscapeDiversityAES Landscape diversity (Shannon index) associated with AES Habitat

LandscapeDiversityNAT Landscape diversity (Shannon index) not associated with Habitat

159

Short name Definition Model

AES

Leg Total surface of legume fields Habitat

LegAES Surface of legume fields included in AES Habitat

LegNAT Surface of legume fields not included in AES Habitat

LegStubbleAES Surface of legume stubble fields included in AES Habitat

LegStubbleNAT Surface of legume stubble fields not included in AES Habitat

LegStubbleTot Total surface of legume stubble fields Habitat

MeanHeight Mean vegetation height in cm Food

Plough Surface of ploughed fields Habitat

Plough2 Surface of ploughed fields with sprouted weeds Habitat

Roughness Roughness of the ground using three categories (low, medium or high)

Food

SeedAES Seed biomass in AES fallows + AES cereal stubble + AES legume + AES legume stubble

Food

SeedCerNAT Seed biomass in non-AES cereal fields Food

SeedCerStubbleAES Seed biomass in AES cereal stubble fields Food

SeedCerStubbleNAT Seed biomass in non-AES cereal stubble fields Food

SeedEdge Seed biomass in edges Food

SeedFallowAES Seed biomass in AES fallow fields Food

SeedFallowNAT Seed biomass in non-AES fallow fields Food

SeedHQFNAT Seed biomass in non-AES ‘high quality fields’ (non-AES fallows + non-AES cereal stubble + non-AES legumes + non-AES legume stubbles)

Food

SeedLegAES Seed biomass in AES legume fields in AES Food

SeedLegNAT Seed biomass in non-AES legume fields Food

SeedLegStubbleAES Seed biomass in AES legume stubble fields Food

SeedLegStubbleNAT Seed biomass in non-AES legume stubble fields Food

SeedNAT Seed biomass in non-AES fields Food

SeedPlough Seed biomass in ploughed fields Food

SeedPlough2 Seed biomass in ploughed fields with sprouted weeds Food

SeedTot Total seed biomass Food

160

Appendix S4. Results of the Principal Component Analyses (PCA) carried out to explore correlations among variables. The most representative variable from

each axis (marked in bold) was selected to be included it in the candidate models for bird abundance, richness, diversity and SPEC-score. The type of model

(food, habitat) to which each variable belongs is shown for the first set of models.

1) Food models

Wintering Variable PC 1 PC 2 PC 3 PC 4 PC 5 PC 6 PC 7 PC 8 PC 9 PC 10 PC 11

ArthrLegStubbleNAT Food 0.018 0.006 -0.029 0.004 -0.039 0.021 0.051 0.047 0.081 -0.968 -0.003

ArthrTot Food 0.041 -0.954 0.014 -0.066 0.011 -0.063 -0.022 0.003 -0.069 0.015 0.001

CerStubbleNAT Habitat -0.034 0.0314 0.057 -0.933 0.005 0.042 -0.049 -0.121 0.068 0.010 0.061

FallowAES Habitat 0.871 0.021 -0.061 0.029 -0.08 -0.016 0.142 -0.066 0.167 -0.132 0.049

LegNAT Habitat -0.063 -0.031 -0.019 0.042 0.023 0.019 0.030 -0.051 0.024 -0.121 -0.958

LegStubbleAES Habitat 0.058 0.004 -0.007 0.016 -0.928 0.031 -0.026 0.004 0.001 -0.065 -0.003

Plough Habitat -0.091 -0.002 -0.132 0.231 0.046 0.122 0.158 0.066 -0.851 0.039 0.091

SeedCer Food 0.033 0.153 0.229 0.084 0.052 0.081 -0.020 -0.768 0.084 0.02 0.144

SeedHQFNAT Food 0.052 -0.129 0.932 -0.102 0.029 0.007 0.110 -0.002 0.099 -0.006 -0.065

SeedLegAES Food 0.055 -0.025 -0.091 0.042 -0.008 -0.937 0.056 -0.031 0.054 0.032 0.066

Mating Variable Model PC 1 PC 2 PC 3 PC 4 PC 5 PC 6 PC 7 PC 8

ArthrNAT Food 0.086 -0.174 0.014 0.486 -0.021 0.084 0.835 -0.061

CerStubbleAES Habitat -0.166 0.015 0.003 -0.064 -0.951 -0.066 -0.036 0.021

FallowNAT Habitat 0.033 -0.249 0.012 0.931 0.085 -0.055 -0.022 -0.034

DifHeight Food -0.115 0.001 -0.936 -0.063 0.145 -0.069 0.018 0.031

LegAES Habitat -0.942 0.034 -0.069 0.015 0.028 -0.031 -0.122 -0.026

LegNAT Habitat 0.068 0.081 -0.049 0.138 0.103 0.081 -0.016 -0.917

Plough2 Habitat 0.165 -0.121 0.006 -0.068 0.057 0.908 -0.006 0.009

SeedFallowNAT Food 0.041 -0.947 -0.011 0.280 0.021 0.038 0.016 0.012

161

Post-fledging Variable Model PC 1 PC 2 PC 3 PC 4 PC 5 PC 6 PC 7 PC 8 PC 9

ArthrFallowNAT Food -0.043 0.921 -0.049 0.018 0.071 0.01 0.003 0.001 -0.006

ArthrLegNAT Food -0.023 -0.02 0.022 0.021 -0.995 0.004 -0.009 -0.03 0.024

CerStubbleAES Habitat 0.288 0.016 0.006 0.072 0.01 0.116 -0.052 0.941 0.001

FallowAES Habitat 0.947 -0.033 0.016 0.063 0.021 -0.037 0.051 0.23 -0.019

LegStubbleAES Habitat 0.261 -0.016 -0.008 -0.103 0.01 -0.972 -0.051 0.001 0.028

LegStubbleNAT Habitat -0.041 0.065 0.021 -0.952 0.01 0.034 0.023 -0.057 0.014

Plough Habitat -0.159 -0.314 -0.912 0.098 0.004 0.006 0.152 -0.089 0.086

Plough2 Habitat 0.036 -0.048 -0.051 0.032 0.041 0.066 0.007 -0.003 -0.943

SeedsLegAES Food 0.169 -0.058 -0.061 0.036 0.039 -0.089 -0.957 0.2162 -0.001

2) Habitat models

Wintering Variable PC 1 PC 2 PC 3 PC 4 PC 5 PC 6 PC 7 PC 8 PC 9 PC 10 PC 11

CerStubbleNAT -0.031 0.341 0.036 -0.933 0.007 0.041 -0.051 -0.112 0.059 0.002 0.061

Edge -0.020 0.018 -0.015 0.061 0.03 -0.044 0.325 0.726 -0.086 -0.043 0.213

FallowAES 0.872 0.014 -0.046 0.039 -0.052 -0.015 0.21 -0.074 0.201 -0.141 0.061

FallowNAT -0.05 -0.723 0.312 0.165 -0.031 -0.052 0.468 0.219 0.044 0.036 -0.019

LegAES 0.351 0.01 0.102 0.002 0.051 -0.831 -0.069 0.171 0.051 0.029 -0.009

LegNAT -0.062 -0.043 -0.036 0.026 0.019 0.019 0.034 -0.036 0.051 -0.113 -0.972

LegStubbleAES 0.054 0.009 -0.008 0.021 -0.934 0.061 -0.022 0.003 0.002 -0.084 -0.002

LegStubbleNAT 0.047 0.012 -0.044 0.026 -0.128 0.039 0.05 0.047 0.071 -0.964 -0.005

Plough -0.122 -0.005 -0.121 0.32 0.041 0.144 0.155 0.057 -0.851 0.062 0.072

SeedFallowNAT 0.058 -0.234 0.876 -0.01 0.003 0.012 0.132 0.119 0.064 0.029 0.011

162

Mating Variable PC 1 PC 2 PC 3 PC 4 PC 5 PC 6 PC 7 PC 8

ArthrCerNAT 0.21 0.121 -0.232 -0.253 -0.068 0.018 0.731 -0.056

CerStubbleAES -0.161 0.015 0.002 -0.063 -0.941 -0.052 -0.032 0.048

FallowNAT 0.034 -0.221 0.012 0.931 0.08 -0.05 -0.019 -0.07

LegAES -0.924 0.014 -0.061 0.011 0.019 -0.024 -0.113 -0.012

LegNAT 0.071 0.049 -0.041 0.131 0.051 0.082 -0.032 -0.916

Plough 0.037 0.18 0.861 -0.166 0.135 -0.158 -0.12 0.063

Plough2 0.151 -0.116 0.04 -0.04 0.062 0.921 -0.007 0.008

SeedFallowNAT 0.026 -0.947 -0.009 0.2 0.018 0.072 0.041 0.03

Post-fledging Variable PC 1 PC 2 PC 3 PC 4 PC 5 PC 6 PC 7 PC 8 PC 9

CerStubbleAES 0.158 0.003 0.004 0.027 0.01 0.018 -0.066 0.928 0.01

FallowAES 0.961 -0.017 0.012 0.051 0.008 -0.035 0.079 0.26 -0.039

FallowNAT -0.05 0.847 -0.069 -0.258 0.081 -0.034 0.069 0.002 -0.019

LegAES 0.406 -0.026 -0.059 0.047 0.061 -0.269 -0.915 0.121 0.011

LegNAT -0.069 -0.038 0.051 0.029 -0.974 0.013 -0.005 -0.061 0.066

LegStubbleAES 0.174 -0.041 -0.012 -0.019 0.006 -0.963 -0.042 0.00 0.039

LegStubbleNAT -0.072 0.109 0.009 -0.957 0.021 0.054 0.044 -0.019 0.013

Plough -0.151 -0.301 -0.897 0.151 0.011 0.007 0.133 -0.047 0.156

Plough2 0.027 -0.045 -0.061 0.049 0.033 0.081 0.013 -0.002 -0.927

163

CAPÍTULO 5

164

Este capítulo se encuentra en fase de preparación:

Ponce, C., Salgado, I., Bravo, C., Gutiérrez, N. & Alonso, J.C. Effects of farming

practices on nesting success of steppe birds in dry cereal farmland.

165

CAPÍTULO 5

Effects of farming practices on nesting success

of steppe birds in dry cereal farmland

Carlos Poncea, Iván Salgadoa, Carolina Bravoa, Natalia Gutiérreza,

Juan Carlos Alonsoa

a Dep. Ecología Evolutiva, Museo Nacional de Ciencias Naturales, CSIC, José

Gutiérrez Abascal 2, E-28006 Madrid, Spain

166

ABSTRACT

Predation is the most common cause of nest failure in ground-nesting birds. Natural

predation rates may be influenced by both, regular agricultural practices and agri-

environmental measures promoted by agri-environmental schemes (AES). These practices

and measures could increase predation rates by increasing predator abundance, or

decrease them by providing more vegetation where birds should find safe nesting places.

We investigated these issues in a dry cereal farmland in central Spain, by means of an

experimental setup with 520 artificial nests of quail. Artificial nests were distributed

among 13 sites, each including all main field types of the area. Predation rate was analyzed

using averaged mixed models, and predictor variables describing the physical

characteristics of the nesting site at three scales (landscape, field, and nest location within

each field - central and peripheral-). Game cameras were used to identify predators and

analyze their nest predation patterns. We found that 6.2% of the nests were destroyed by

tractors ploughing in spring. Overall nest predation rate was 66.3%, affecting more to

nests surrounded by organic cereal crops and ploughed fields. Nests located near field

margins suffered more predation than those in the centre of the fields. Nest predation was

highest in ploughed fields, intermediate in AES-promoted fields (fallows, organic cereal,

vetch), and lowest in regular cereal fields. In all field types nest predation rate decreased

with increasing vegetation height, because tall vegetation offered good nest concealment

opportunities. Fallows, vetch fields and organic cereal fields provided intermediate

vegetation heights, and thus relatively safe nesting sites. However, but due to the high food

availability predators could find on these substrates, they acted as ecological traps

because predators concentrated their predation activity on them. Camera traps recorded

42 predation events (81% of birds, 19% of mammals). The main predators were Marsh

Harrier, Montagu's Harrier and Common Buzzard. Spring ploughing should be restricted

to prevent nest destruction. Fields promoted by AES should be dispersed in order to avoid

attracting nest predators.

Keywords

Agri-environment scheme, farmland, ground-nesting bird, habitat management,

predation .

167

INTRODUCTION

Dry cereal farmland holds important breeding and wintering populations of many

farmland bird species of European conservation concern (Butler et al. 2010).

Although agriculture has favoured the expansion of farmland habitats in past

centuries, the replacement of traditional by intensive farming practices in recent

times has led to habitat changes whose negative effects have been highlighted in

numerous studies. The main practices associated to intensive agriculture are land

management changes (e.g., moving the ploughing of cereal stubbles forward by

several months, from spring to immediately after harvest), loss of crop diversity,

increased pesticide and fertilizer use, removal of edges and uncultivated areas,

and earlier harvest dates (Newton 2004). All these changes have resulted in the

loss of suitable feeding and nesting habitats, and a reduction in food and nesting

places available. One of the main consequences of this habitat deterioration has

been a significant decline suffered by European farmland bird populations during

the last decades (Donald et al. 2001, Gregory et al. 2005). Like most bird

populations in northern and central Europe, those of the Mediterranean region

have also suffered the consequences of recent changes in farmland habitats. For

example, according to the last Common Breeding Bird Monitoring Scheme report

in Spain (SACRE), the numbers of farmland breeding birds have declined by 25%

during the last 17 years, and almost 30% inside Important Bird Areas (Escandell

2015).

Predation has been identified as the most important cause of nest failure of

ground-nesting birds in farmland habitats (Draycott et al. 2008), and thus a

relevant factor determining their decline (Donald et al. 2002, Bradbury et al.

2000). High predation rates may limit the breeding populations of farmland

species (Gibbons et al. 2007), or influence their demographic parameters

(Whittingham & Evans 2004). Thus, an interesting research issue is how intensive

agricultural practices interact with nest predation. It is known that intensive

agriculture leads to increasing predation rates in farmland species (Tapper et al.

1996, Paridis et al. 2000, Newton 2004). At least three hypotheses have been

proposed to explain this phenomenon. The first is that high predation rates could

168

be the result of a higher density of predators (Baillie et al. 2002, 2007, Evans

2004). A second hypothesis suggests that predation rates can increase

independently of changes in predator density (Wilson et al. 1997a, Donald 1999,

Pescador & Peris 2001). For example, a decrease in suitable habitat availability can

make prey more vulnerable to predation (Whittingham & Evans 2004) or force

some species to concentrate their nests in the remaining smaller patches of

appropriate habitat, which would suffer higher predation pressure and thus turn

into "ecological traps" (Chamberlain et al. 1995, Pescador & Peris 2001). Also,

predators can shift their diet and select new prey if their usual target species have

declined as a consequence of agricultural intensification (Schmidt 1999, Evans

2004, Newton 2004,). In fact, some of the characteristics of agricultural

intensification are the destruction of edges and fallows, or the homogenization of

crops, leaving only small isles of suitable habitat for nesting. A third hypothesis

suggests that if removing edges and fallows forces ground-nesting species to nest

in more exposed places, nest predation will increase. Whatever the reason, the link

between increasing nest predation rates and agricultural intensification seems

clear.To reverse the decline of farmland birds, agri-environmental schemes (AES)

have been running in many countries over the last decades (reviewed in Kleijn &

Sutherland 2003). Most AES include payments to farmers for implementing

measures that benefit wildlife (Kleijn & Sutherland 2003). While studies evaluating

and proposing actions intended to increase food availability are relatively

abundant (Campbell et al. 1997, Herkert 2009, Lapiedra et al. 2011), those relating

nesting habitat quality and predation rates are less common (but see Beja et al.

2013, Evans 2004, Fletcher et al. 2010).

Increasing natural vegetation (e.g., by creating fallows) and landscape

heterogeneity (e.g., by introducing different crops) are some of the measures

usually proposed in AES (Kleijn & Sutherland 2003). Besides providing food

resources, these measures contribute to reduce nest predation rates by providing

safe nesting places (Newton 1998, Wilson et al. 2001a, Beja et al. 2013), or by

increasing the number of potential nest sites that predators must search, thus

reducing nest density in each field and improving nest concealment (Bowman &

Hams 1980, Martin 1993). But increasing high quality habitats may also attract

predators (Pescador & Peris 2001).

169

Although nest concealment is an important factor influencing predation rate

on farmland habitats (Yanes & Suárez 1995, Magaña et al. 2010), the type of

predator is also relevant (Patten & Bolger 2003. Bayne et al. 1997). Therefore,

understanding how the foraging patterns of all potential predators relate to

nesting sites, landscape and habitat characteristics is crucial for implementing

appropriate management actions (Benson et al. 2010). Also, birds use visual

stimuli for nest detection, whereas mammals use the sense of smell and thus, nest

concealment should not be a crucial factor in mammal nest predation avoidance

(Rangen et al. 1999). Studying the predator community seems therefore essential

to understand the reasons of the success or failure of AES involving nest

concealment.

In this study we assessed nest predation in relation to both, ordinary

agricultural practices applied in Mediterranean dry cereal farmland, as well as

additional measures from a currently running AES, in a farmland area of central

Spain. We used an experimental design with artificial nests, and game cameras to

detect and identify predators. We also considered the influence of nest location

within the field and the characteristics of the surrounding habitat (landscape

context), since these factors can modify nest predation (Storch et al. 2005, Reino et

al. 2010). Our hypothesis was that nest predation would be affected by the micro-

habitat structure around nests, the landscape characteristics, and the differences in

foraging patterns among predators. Specifically, we predicted that increasing

vegetation height and cover, and landscape diversity derived from AES measures

would reduce predation rates by favouring nest concealment and providing

suitable nesting places. We also predicted that predation rates would be higher on

edges, which are commonly used by some predators (Blouin-Demers &

Weatherhead 2001). In light of the results, some management actions that could

contribute to reduce nest predation are discussed.

170

METHODS

Study area, field selection and nest placement

The study was carried out in the Special Protected Area (SPA) 139 "Estepas

cerealistas de los ríos Jarama y Henares", located in the north-eastern part of

Madrid province (central Spain), where an agri-environmental scheme (AES) has

been running since 2003. Briefly, the AES consisted on payments to farmers for

growing organically vetch and cereal, interruption of the cereal production (long-

term fallows) and maintenance of cereal stubble during the winter (for more

details on AES measures see Ponce et al. 2014). The landscape in the SPA is

homogeneous, and the area is mainly dedicated to dry cereal cultivation (wheat

Triticum aestivum, barley Hordeum vulgare, and oat Avena sativa) with some

dispersed bushes (Retama sphaerocarpa, Thymus spp., etc.) and sporadic trees

(Quercus ilex, Pinus spp.). Some of the most common ground-nesting birds in the

SPA are Calandra Lark (Melanocorypha calandra), Corn Bunting (Miliaria

calandra), Great Bustard (Otis tarda), Crested Lark (Galerida cristata) and Little

Bustard (Tetrax tetrax) (Ponce et al. 2014), all of them included in the list of

Species of European Conservation Concern (BirdLife International 2004).

There is no consensus regarding the applicability of predation rates

obtained in experimental studies with artificial nests to natural situations (Martin

1987 and Matessi & Bogliani 1999, Mezquida & Marone 2003, Robinson et al. 2005,

Noske et al. 2008). However, artificial nests are considered useful for identifying

factors affecting spatial and temporal predation patterns (Major & Kendall 1996,

Roos 2002, Batary & Baldi 2005, Ludwig et al. 2012, Mandema et al. 2013) in

comparative studies and different settings. Also, nest predation of artificial nests

seems to be useful for establishing relative predation pressures, at least for

ground-nesting passerines in open habitats with low structural complexity, where

rates from experimental and natural situations were found to be similar (Vögeli et

al. 2011). Furthermore, artificial nests allow using sample sizes in habitats that

may be avoided by birds and controlling the parameters to be studied (Beja et al.

2013).

171

To place artificial nests we selected 13 sites in the SPA (Figure 1). Each of

these 13 sites included all kind of fields from the AES (long-term fallows, vetch and

organically cultivated cereal ) and those from regular agricultural activities

(ploughed fields and cereal crops). We were particularly interested in investigating

the effects of fields promoted by AES (fallows, vetch and organic cereal) on nest

predation rate. We selected fields as close together as possible to maximize the

probability that all fields of that site were located within the territory of a given

predator, and thus the predation probability was the same among fields within a

given site.

Figure 1. Location of the study area in the Iberian Peninsula. The figure on the right shows the

Special Protected Area (SPA) 139 "Estepas cerealistas de los ríos Jarama y Henares" and the 13 sites

where the field work was carried out.

We carried out the experiment in two trials during the spring in 2012.

Artificial nests were placed on 16th May for the first trial and on 5th June for the

second trial. We monitored nests during 14 days which is equivalent to a nesting

cycle of most passerine species present in the study area. We removed non-

predated nests at the end of each trial. We also left 1 week between the end of the

first trial and the start of the next one, to avoid any habituation of predators. We

placed 4 artificial nests in each field totalling 520 nests (260 for each trial, and 20

nests at each site). Within a field, we placed 2 nests close to the edge (at 2 m) and 2

172

in well inside the field (at 50 m from the edge nest towards the centre of the field;

this distance is higher than those considered in most studies, see e.g. Díaz &

Carrascal 2006). This design minimizes the probability that two nests of the same

field are found by the same predator, and thus maximizes independence of

predation results among nests. Each artificial nest consisted of a slight depression

on the ground to avoid their displacement, with 3 fresh Quail (Coturnix sp.) eggs.

We used eggs bought at industrial quail farms because they are easy to obtain in

large quantities and similar to those that can be found naturally in farmland areas

(de Graaf & Maier 1996, Maier & de Graaf 2001). We used eggs from two different

companies, but showing no differences in length, width and weight (n=150, p

>0.43 in all cases). Close to each nest we placed a reed (1.20 m height) to enable an

easy location of the nest in subsequent visits. A small tag with information about

the study was attached to the top end of the reed.

Since predators were identified by means of game cameras (see below), we

also tested the possible effect of these cameras on nests predation (Richardson et

al. 2009). We monitored 41 nests with game cameras and compared them with a

sample of 46 control nests without game cameras. All of them were visited 3 times

(to avoid the "researcher effect") evenly distributed along the study periods, and

after cameras were installed. We did not find any effect of placing game cameras

on artificial nests predation (Chi-Square = 0.01, P= 0.93). Since some studies have

shown that increasing the number of visits has an effect on nest predation (e.g.,

Verboven et al. 2001, but see Ibáñez-Álamo et al. 2012) we also tested the

researcher effect. We visited 46 nests 3 times and 46 only 1 (plus the placement

day in both samples). We did not find any effect of the number of visits (Chi-Square

= 0.21, P= 0.65). Accordingly, all artificial nests were included in subsequent

analyses.

Field and landscape variables description

We distinguished three groups of variables: those related to the micro-habitat

structure around the nest (measured within a circle of 2 m diameter centred on the

nest); those related to the nest location, and those describing the landscape

context, measured within a circle of 100 m radius around the nest, since it is

known that landscape composition and structure may affect predator composition

173

and abundance (Pita et al. 2009). Prior to starting the nest visits, all observers

participating in the fieldwork spent one day standardizing the measures and

estimates to be taken.

For each nest we recorded the UTM coordinates using a GPS (Garmin, ±2 m

error). Vegetation abundance and structure were assessed visually around the

nest. Data recorded were total plant cover (% of the surface) and vegetation height

(maximum, mean and most frequent, in cm). Nest location variables were the field

type (ploughed, cereal crop, organic cereal crop, vetch and fallow) and the location

in the field (edge or interior). Landscape variables were distances (in m) to the

nearest shrub or tree (estimated visually), and watercourses and paths, presence

of a watercourse, length of watercourses and edges (in m), and the surface of each

type of field plus the surface of buildings (including farms) and shrubs (based on

GIS information).

Video monitoring

We installed 21 game cameras (model Bresser 3310000) at random nests trying

that they were evenly distributed among sites, field types, and location in the field

(edge, interior). The aim was to record any predation event and identifying the

whole range of potential predators. Game cameras allow identifying nest

predators, in contrast to tracks or faeces near the nest (Benson et al. 2010).

Game cameras were sensitive to any movement around the nest and

recorded 1 minute videos after a movement was detected during day (coloured

videos) or night (by infrared illumination, black and white videos). After the first

minute was recorded the camera started recording a new video until the predator

had gone away or the memory card was full. Cameras were placed at a distance of

ca. 50 cm from the nest and ±50 cm height above the ground by attaching them to a

stick with a small tag providing information about the study. We reviewed each

camera at 4 to 5-days intervals to change batteries and memory cards until a nest

failed or was successful (14 days). When a nest failed the camera was moved to

another nest. In total, we used 132 cameras, recording 4137 videos, each of 1-

minute duration.

174

Data analyses

The sample unit for this study was the nest, and the response variable was nest

predation (yes or no) at the 14th day after nest placement. A nest was considered

predated when at least one of the eggs showed any break in the shell or had

disappeared or was clearly displaced from the nest (Mezquida & Marone 2003,

Batary et al. 2004). Five nests not found during the visits for unknown reasons

were excluded from the analyses. All percentage values (%of field types and

vegetation cover) were arcsine square root transformed.

To study the most important variables influencing nest predation we built

averaged mixed models. We first selected candidate continuous variables by using

principal component analysis (PCA) with the Varimax Normalized factor rotation.

The minimum eigenvalue was 1. We selected the variable with the higher

correlation value from each axis (Table 1) to reduce multicollinearity (Beja et al.

2013, Dormann et al. 2013, Barrientos & Arroyo 2014, Ponce et al. 2014).

175

Table 1. Results of the Principal Component Analysis (PCA) performed on the whole set of

continuous variables. The most representative variable from each axis (marked in bold) was

selected to be included in the candidate models for nest predation.

Variable PC 1 PC 2 PC 3 PC 4 PC 5 PC 6 PC 7 PC 8 PC 9

Vegetation cover 0.73 0.08 0.02 -0.50 -0.02 -0.07 0.00 0.05 -0.18

Distance to the nearest edge -0.01 -0.01 0.01 -0.05 -0.97 -0.01 -0.01 -0.04 0.01

Distance to the nearest path 0.16 0.14 0.06 0.02 -0.05 -0.02 -0.02 -0.11 0.05

Distance to the nearest river 0.03 -0.02 0.95 0.02 -0.03 0.02 -0.01 -0.02 -0.14

Length of edges 0.00 0.00 -0.09 0.19 0.09 0.08 -0.01 0.02 -0.05

Landscape diversity 0.08 0.32 -0.13 -0.23 0.39 -0.06 0.07 -0.24 0.17

Maximum vegetation height 0.89 0.04 0.04 -0.17 0.00 -0.05 0.19 -0.02 0.01

Mean vegetation height 0.94 0.02 0.06 0.14 0.02 0.06 0.05 -0.10 -0.16

Median vegetation height 0.87 0.03 0.05 0.25 0.03 0.11 -0.12 -0.12 -0.22

Length of watercourses -0.08 0.03 -0.93 -0.01 -0.01 0.01 -0.01 0.05 -0.13

Surface of cereal crop 0.25 -0.43 0.02 0.21 0.06 0.10 -0.04 0.29 -0.73

Surface of fallows -0.03 -0.15 -0.03 -0.93 -0.03 0.07 0.03 0.15 -0.03

Surface of organic vetch 0.11 0.96 -0.04 0.15 0.03 0.01 -0.05 0.05 0.05

Surface of organic cereal crop

0.15 -0.03 0.08 0.16 -0.02 0.06 -0.04 -0.96 0.01

Surface of ploughed -0.43 -0.14 0.00 0.29 0.05 0.05 0.03 0.19 0.80

Surface of shrubs -0.04 0.00 -0.01 0.05 0.00 -0.99 0.00 0.05 0.02

Surface of urbanized areas 0.09 -0.03 0.00 -0.02 0.02 0.00 0.99 0.03 0.03

We built the "beyond optimal model" (Zuur et al. 2009) with variables from

the PCA plus categorical variables (location within the field -edge, interior- and

field type) and different random factor structure. The error structure was binomial

for the response variable. As plausible random factors we considered the site, trial

and trial nested within group. We used the results from the ANOVA test in R (R

Development Core Team 2013) to select the best random structure. The random

structure selected was the trial nested within site. We built all possible models and

selected those with an increase of<5 in the Akaike´s Information Criterion (ΔAIC)

over the best model as candidate models (Burnham & Anderson 2002). Finally, we

performed an average model estimation with the package MumIN (Barton 2013) in

R. The final averaged model included those variables with a significant effects on

the response variable: those whose confidence limits excluded zero, since they

have no equivocal meaning (Delgado et al. 2013, Ponce et al. 2014, Beja et al.

2013).

176

Predator behaviour was tested by means of Chi square test. We tested which group

of predators was more frequently recorded. We also tested if there were different

predation patterns in relation to the field type and the location of the camera in the

field (edge or interior).

RESULTS

We found 32 nests ploughed or run over by tractors. Most of them were placed in

ploughed fields but there were also some nests destroyed by tractors in organic

vetch crops, long-term fallows, and organic cereal crops.

We found 320 nests predated (66.3%) at the end of the study. Model

averaging showed that predation was influenced by the surface of organic cereal

crop and the surface of ploughed fields around the nest, the type of field where

nests were placed, the location of the nest in the field, and the mean height of the

surrounding vegetation (Table 2).

177

Table 2. Model-averaged estimates for nest predation in a dry cereal farmland area (Special Protected Area for Birds no. 139) in central Spain. The statistics

given are: sum of Akaike weights of the models in which the predictor was retained (Σ), parameter estimate of the regression equation (b), standard

deviation of the regression parameter (SE), lower and upper confident limits of b (Lower, Upper CI), and standardized coefficients of predictors (β).

Significant predictors are marked in bold.

Scale Predictor ∑ b SE Z value P Lower CI Upper CI β

Intercept

2.23 0.68 3.30 0.001 0.88 3.59 0.15

Landscape Surface of organic cereal crop 1.00 2.26 0.72 3.17 0.002 0.83 3.69 0.16

Surface of ploughed 1.00 2.21 0.50 4.42 0.000 1.21 3.21 0.11

Surface of schrubs 0.47 1.52 1.28 1.19 0.234 -1.04 4.07 0.19

Surface of fallows 0.23 0.42 0.69 0.60 0.549 -0.97 1.80 0.03

Surface of organic vetch 0.29 -0.59 0.64 0.92 0.356 -1.87 0.69 -0.04

Surface of urbanized areas 0.26 -0.26 2.47 0.11 0.915 -5.21 4.68 -0.06

Distance to the nearest river 0.42 0.00 0.00 1.34 0.180 -1.5E-03 2.9E-04 -2.5E-08

Distance to the nearest edge 0.29 0.00 0.01 0.36 0.721 -0.02 0.02 -3.0E-06

Field Ploughed 0.97 -2.42 0.70 3.46 0.001 -3.82 -1.02 -0.16

Organic cereal crop 0.97 -1.41 0.68 2.09 0.037 -2.76 -0.06 -0.09

Cereal crop 0.97 -0.92 0.44 2.09 0.037 -1.80 -0.04 -0.04

Location in the field (interior) 0.81 -0.55 0.27 2.07 0.039 -1.09 -0.02 -0.01

Organic vetch 0.97 -0.63 0.54 1.18 0.239 -1.70 0.44 -0.03

Nest Mean vegetation height 1.00 -0.03 0.01 4.33 0.000 -0.05 -0.02 -2.5E-05

178

The model predicts that the risk of predation increased in nests surrounded

by more surface of organic cereal or ploughed fields. Also, predation was higher in

fields with lower vegetation and in nests located near the edge of the field (70.3%,

compared to a 62.3% in nests placed inside fields). The highest predation rate

observed was for ploughed fields (83.3%), followed by long-term fallows (76.5%),

organic cereal crops (69.9%) and vetch crops (61.4%), whereas regular cereal

crops showed the lowest predation rate (43.4%, Figure 2).

Figure 2. Mean predation rate in the different field types.

We recorded a total of 42 predation events belonging to seven wild animal

species, and feral dogs. We also recorded some potential predators as feral cats,

Montpellier snake (Malpolon monspessulanus) and rats (Rattus spp.) crossing in

front of the camera, but none of these predated any nest. Birds were involved in

more predation events than mammals (respectively, 81% and 19%, Chi-Square =

16.1, p < 0.01). The main group of bird predators were raptors. We recorded three

raptor species, the Western Marsh Harrier (Circus aeruginosus), Montagu's Harrier

(C. pygargus), and Common Buzzard (Buteo buteo), which altogether preyed upon

28 nests. Six nests were predated by Magpies (Pica pica). The most common

mammal detected was the Stone Marten (Martes foina), which predated on three

nests, followed by feral dogs and Edgehogs (Erinaceus europaeus) with two

179

predation events each, and Western Mediterranean Mouse (Mus spretus), which

preyed upon one nest.

Birds did not differ on the type of field where they predated (Chi-Square =

4.8, p = 0.31), but mammals did it more intensively on fallows (Chi-Square = 11.7, p

= 0.02). Also, mammals predated more intensively on nests placed on the edge

than in the interior of the field (Chi-Square = 5.1, p = 0.02). There was no

differential predation patterns for bird species (Chi-Square = 0.46, p = 0.50) or for

raptors alone (Chi-Square = 0.18, P = 0.67).

DISCUSSION

Thirty two (6.2%) out of a total of 520 nests used in this experiment were

destroyed or run over by tractors during their labours. This seems to be a common

problem in farmland areas worldwide (Newton 2004, Tews et al. 2013). Sánchez-

Oliver et al. (2014) and Reino et al. (2010) reported that respectively, 12% and

4.5% of their nests were also damaged by ploughing activities. Although most

birds do not select ploughed fields for nesting, some species such as larks, curlews

and lapwings prefer them over other substrates (Berg et al. 1992, Wilson et al.

1997b, Galbraith 1988, Green et al. 2000). Farmers plough their fields several

times along the year, and farming activities are especially common in intensive

agriculture areas during spring. In our study area, the common cycle of ploughing

operations usually starts after harvesting in early summer, when around 20%-

30% of the fields are ploughed. A second ploughing period occurs during the

winter (80%-100% of the fields) and another period follows during the next spring

or summer. These fields are sown in the following autumn or winter after a new

ploughing operation. Although the objectives of such practices are to avoid

nutrients and water loss due to the growth of weeds, the main consequence for

farmland species is a marked decrease in suitable nesting habitat (Berg et al.

1992). However, the most common weeds found in our area during the normal

nesting period (Salsola kali, Heliotropium europaeum, Solanum nigrum, Datura

stramonium, etc.) develop their biological cycle in late spring and summer

(Villarías et al. 2006), and thus do not really compete with cereal plants during

180

their growing or seed maturation periods. It is well known that ploughed fields

with sprouted weeds are important for farmland birds during the breeding and

post-fledging periods (Ponce et al. 2014). For that reason, we consider that current

weed control during winter and early spring is enough to prevent cereal

production being affected by any weed pest. We suggest that ploughing labours

and, in general, field operations should be significantly reduced during this period

to avoid the destruction of nests (Wilson et al. 2005).

The mean nest predation rate recorded in this study was 66.3%. This is an

intermediate figure compared to other studies using artificial nests in the Iberian

Peninsula. Sánchez-Oliver et al. (2014) found rates of 88.4% in open farmland

habitat, though using more days of exposure, whereas Reino et al. (2010) obtained

49% predation rate after 15 days of exposure. Pescador & Peris (2001) found 61%

predated nests after 15 days of exposure in the field.

Our results showed that different parameters related to landscape

characteristics, nest location, and micro-habitat structure around the nest

influenced nest predation patterns. Landscape composition played an important

role, as suggested in previous studies (Reino et al. 2010, but see Beja et al. 2013).

Most studies relating nest predation and landscape features compared forest

plantations with open areas (e.g., Batáry & Báldi 2004), but here we have found

differences within a relatively uniform landscape, the dry cereal farmland. The

influential variables were the surface of organic cereal crop and ploughed fields

around the nest, which showed the highest beta values in the model. Both variables

had the same effect on nest predation. Predated nests had higher proportions of

these two field types in a surface of 100 meters radius around the nest. However,

both field types suggest contradictory explanations, since the amount of vegetation

differs much between them: organic cereal fields have abundant vegetation

whereas ploughed fields have very scarce or no vegetation at all. The most

plausible explanation is that higher predation rates in these fields increase the

likelihood of predation in surrounded fields (Wilson et al. 2001b). Also, predators

include high quality fields (as organic cereal crops) in their home ranges, and fields

joined to organic crops can attract predators (Reino et al. 2010).

181

Predation patterns also varied at the field scale, differing among specific

field types in which the nest was located. The highest predation rate was found in

ploughed fields, where the absence of sufficient vegetation makes nests more

vulnerable. It is necessary to highlight that some ground nesting species like

curlews and skylarks typically select fields with scarce vegetation for nesting

(Green et al. 2000, Whittingham et al. 2002), probably because this enables them

to detect approaching predators and adopt anti-predator behaviours such as

distraction displays (Evans 2004). Surprisingly, nests located in long-term fallows

and organic cereal crops were also highly preyed upon. Many studies have pointed

out the importance of uncultivated and organic crop fields for biodiversity and for

bird nesting (O’Connor & Shrubb 1986, Wilson & Browne 1993, Brickle et al. 2000,

Hole et al. 2005). In our study, both field types were favoured by the AES, and thus

agri-environmental measures did not succeed in protecting nests from high

predation rates, in spite of theoretically providing more nest concealment

opportunities. As explained above for the landscape structure in circles of 100 m

radius around nesting sites, fallows and organic cereal fields probably attracted

predators due to the higher prey abundance and diversity contained in them, and

so the probability of nests being preyed upon increased in these substrates.

Vegetation height was a significant factor in the model describing predation

probability, but the vegetation grown in fallows and organic fields did not reach a

height sufficient to offer optimal nest concealment opportunities. The same result

was reported by Pescador & Peris (2001), who also suggested that fallows

attracted predators. The same result was reported by Pescador & Peris (2001) in

Spain. They suggest that fallows are scarce in their study area and predators are

attracted to them. In the place where we carried out the experiment both fields

were present. However, they are scarce at higher scale.

Our results support the idea of "ecological trap" suggested by other

researchers to explain high predation rates in high quality fields (Chamberlain et

al. 1995, Donald 1999, Evans 2004, Sokos et al. 2013). It is well known that long-

term fallows and organic cereal crops are highly selected by birds during the

nesting period (Beja et al. 2013, Newton 1998, Wilson et al. 2001a, Hole et al.

2005). A balanced pay-off might exist between the high amount and quality of food

available in these fields and the high predation risk on them. Nests in fallows can

182

suffer relatively high predation rates, but on the positive side, the abundant food

available on them allows chicks to grow and survive better compared to other

fields (Hole et al. 2005).

The lowest predation rates were detected in regular cereal crops. These

fields had the tallest vegetation, and thus offered the best concealment

opportunities for nesting. In our study area, cereal crops were selected by Great

bustards as preferred substrates for nesting (Magaña et al. 2010). Other authors

have also found that nest concealment may decrease the probability of nest

predation (Rangen et al. 1999, Beja et al. 2013). Besides, in a natural situation the

density of nests is probably lower in cereal fields than in other substrates (Wilson

et al.1997b), due to the low food availability, which plays an important role for

nest site selection (Kragten & De Snoo 2007, Kragten et al. 2008).

Most of previous nest predation studies were carried out comparing edges

with forested areas (e.g., Benson et al. 2010). In our study, nests placed near the

field edge were more intensively preyed upon than those located inside the fields.

There are different plausible reasons for this result. First, field edges are

frequently used by birds during the breeding season for nesting, which makes

them attractive and profitable sites to predators searching for food (Gates & Gysel

1978, Chalfoun et al. 2002). Second, it is known that predators use linear

structures to move between areas (Bider 1968), although this behaviour is more

common in mammals tan in birds.

Finally, camera trapping results showed that different predators selected

different habitats to search for food (Bayne et al. 1997, Benson et al. 2010). In our

study area cameras detected four times more birds than mammals, and the most

frequent bird predators were raptors. Birds, and particularly raptors are well

known nest predators (Opermanis 2001, Batary et al. 2004, King & DeGraaf 2006,

Purger et al. 2008). Benson et al. (2010) found that raptors concentrated their

searching effort near field margins. In our study, neither all birds nor the group of

raptors showed any preference for field margins. In contrast, mammals showed a

higher predation effort near edges than inside fields. Food searching patterns

differ widely between mammals, which use olfactory cues and typically follow the

same routes like paths, field borders, or their own previous tracks, and birds,

183

which search for nests by means of visual stimuli and cross fields flying without

obvious obstacles (Rangen et al. 1999).

MANAGEMENT IMPLICATIONS

Nest destruction by tractors that plough fields during the breeding period is a

problem in our study area. Since other studies have also identified this problem,

we believe it is important that future AES consider including some restrictions on

agricultural machinery to prevent direct nest destruction. Measures should also

allow some vegetation growth to favour birds nesting in ploughed fields (Donald et

al. 2002).

Predation is considered one of the causes of the recent declines observed in

ground-nesting populations of farmland species. The origin of this high predation

is related to agricultural intensification (Newton 2004). The impact of predators

on farmland species can be counteracted in different ways. One possibility is

predator control (Suárez et al. 1993), which has been proved effective for several

species (Thirgood et al. 2000, Fletcher 2006). However, this possibility does not

solve the problem in the long term. Also, in our study the main predators were

raptors, which are strictly protected by national and international law. A second

possibility is through habitat management tools (Evans 2004). In this case, the

correct design of AES is essential for reducing nest predation of ground-nesting

species.

The results from our study provide strong evidences that predation rates

are influenced by factors acting at landscape, field and nest-site scale. At the

landscape scale, we found predation increases for two situations strongly differing

in vegetation cover around the nest site: areas where the surface of organic cereal

were dominant, and those where ploughed land was dominant. This was

corroborated at the field scale, where nest predation was higher in fallows and

organic cereal crops, two of the field types promoted by AES. One possible way to

minimize this predation increase would be to disperse these fields promoted by

AES in order to avoid the development of isles of high quality, where predators are

expected to focus their hunting efforts. In this way, the food-related benefits of

184

agri-environmental fields would be the same, but would not be counteracted by

high predation rates. In addition, to prevent a higher predation rate on nests near

field borders, we suggest that fields from AES should be large, allowing birds to

find appropriate nesting sites far from edges (Bayne et al. 1997).

Finally, studying the composition of the predator community is necessary to

understand the mechanisms underlying predation rates in different scenarios

(Benson et al. 2010). In areas where birds are the main nest predators, AES

measures should focus on increasing vegetation height, to maximize the offer of

nesting sites with appropriate concealment against aerial predators (Davis 2005).

All these measures aiming at reducing nest predation should be taken into account

together with measures enhancing food abundance when designing AES programs

in dry cereal farmland areas.

AKNOWLEDGEMENTS

Authors want to acknowledge Iris Calleja for her help during field work. We also

thank all farmers in the study area for their collaboration. Compensatory payments

to farmers were financed through an AES funded by the construction of the

highway Madrid-Gadalajara. C.P. was supported by a contract CSIC-HENARSA. The

study was financed by the General Directorate for Scientific Research of the

Spanish Ministry for Science and Innovation (projects CGL2005-04893 and

CGL2008-02567).

REFERENCES

Baillie, S.R., Crick H.Q.P., Balmer, D.E., Beaven, L.P., Downie, I.S., Freeman, S.N.,

Leech, D.I., Marchant, J.H., Noble, D.G., Raven, M.J., Simpkin, A.P., Thewlis,

R.M. & Wernham, C.V. 2002. Breeding Birds in the Wider Countryside: Their

Conservation Status 2001. BTO Research Report no. 278. Thetford: BTO.

Baillie, S.R., Marchant, J.H., Crick, H.Q.P., Noble, D.G., Balmer, D.E., Barimore, C.,

Coombes, R.H., Downie, I.S., Freeman, S.N., Joys, A.C., Leech, D.I., Raven, M.J.,

185

Robinson, R.A. & Thewlis, R.M. 2007. Breeding Birds in the Wider

Countryside: Their Conservation Status 2006. BTO Research Report No.

470. Thetford: BTO.

Barrientos, R., & Arroyo, B. 2014. Nesting habitat selection of Mediterranean

raptors in managed pinewoods: searching for common patterns to derive

conservation recommendations. Bird Conservation International: 24: 138-

151.

Barton, K. 2013. MuMIn: Multi-model inference. R package version 1.9.13.

Batáry, P., Winkler, H. & Báldi, A. 2004. Experiments with artificial nests on

predation in reed habitats. Journal of Ornithology 145: 59-63.

Bayne, E.M., Hobson, K.A. & Fargey, P. 1997. Predation on artificial nests in relation

to forest type: contrasting the use of quail and plasticine eggs. Ecography

20: 233-239.

Beja, P., Schindler,S., Santana, J., Porto, M., Morgado, R., Moreira, F., Pita, R., Mira, A.

& Reino, L. 2013. Predators and livestock reduce bird nest survival in

intensive Mediterranean farmland. European Journal of Wildlife Research:

60: 249-258.

Benson, T.J., Brown, J.D., & Bednarz, J.C. 2010. Identifying predators clarifies

predictors of nest success in a temperate passerine. Journal of Animal

Ecology 79: 225-234.

Bider, J.R. 1968. Animal activity in uncontrolled terrestrial communities as

determined by a sand transect technique. Ecological Monographs 38: 269–

308.

Blouin-Demers, G., & Weatherhead, P.J. 2001. Habitat use by black rat snakes

(Elaphe obsoleta obsoleta) in fragmented forests. Ecology: 82, 2882–2896.

Bradbury, R.B., Kyrkos, A., Morris, A.J., Clark, S.C., Perkins, A.J. & Wilson, J.D. 2000.

Habitat associations and breeding success of yellowhammers on lowland

farmland. Journal of Applied Ecology 37: 789–805.

186

Brickle, N.W., Harper, D.G.C., Aebischer, N.J. & Cockayne, S.H. 2000. Effects of

agricultural intensification on the breeding success of corn

buntings Miliaria calandra. Journal of Applied Ecology 37: 742–755.

Burnham, K.P.,& Anderson, D.R. 2002. Model selection and multi-model inference.

A practical information-theoretic approach. Springer, New York.

Butler, S. J., Boccaccio, L., Gregory, L.D., Vorisek, P. & Norris, K. 2010. Quantifying

the impact of land-use change to European farmland bird populations.

Agriculture, Ecosystems and Environment 137: 348–357.

Campbell, L. H., Avery M.J., Donald P., Evans A.D., Green R.E., & Wilson J.D. 1997. A

review of the indirect effects of pesticides on birds. JNCC Report No. 227.

JNCC, Peterborough.

Chalfoun, A.D., Ratnaswamy, M.J. & Thompson, F.R. III. 2002. Songbird nest

predators in forest-pasture edge and forest interior in a fragmented

landscape. Ecological Applications 12: 858–867.

Davis, S.K. 2005. Nest-site selection patterns and the influence of vegetation on

nest survival of mixed-grass prairie passerines. Condor 107: 605-616.

Delgado, M.P., Sanza, M., Morales, M.B., Traba, J., & Rivera, D. 2013. Habitat

selection and coexistence in wintering passerine steppe birds. Journal of

Ornithology 154: 469–479.

Díaz, L. & Carrascal, L.M. 2006. Influence of habitat structure and nest site features

on predation pressure of artificial nests in Mediterranean oak forests.

Aredeola 53(1): 69-81.

Donald, P.F. 1999. The Ecology and Conservation of Skylarks Alauda arvensis on

Lowland Farmland. DPhil thesis, Oxford University.

Donald, P. F., Green, R.E., & Heath., M.F. 2001. Agricultural intensification and the

collapse of Europe’s farmland bird populations. Proceedings of the Royal

Society of London. Series B 268: 25–29.

187

Donald, P.F., Evans, A.D., Muirhead, L.B., Buckingham, D.L., Kirby, W.B. & Schmitt,

S.I.A. 2002. Survival rates, causes of failure and productivity of Skylark

Alauda arvensis nests on lowland farmland. Ibis 144: 652–664.

Dormann C.F., Elith J., Bacher S., Buchmann C., Carl G., Carré G., García Marquéz J.R.,

Gruber B., Lafourcade B., Leitão P.J., Münkemüller T., McClean C., Osborne

P.E., Reineking B., Schröder B., Skidmore A.K., Zurell D., & Lautenbach S.

2013. Collinearity: a review of methods to deal with it and a simulation

study evaluating their performance. Ecography 36: 27-46.

Draycott, R.A.H., Hoodless, A.N., Woodburn, M.I.A. & Sage, R.B. 2008. Nest predation

of European Common Pheasants Phasianus colchicus. Ibis 150 (Suppl. 1):

37–44

Escandell, V. 2015. Programa Sacre. En, SEO/BirdLife. Programas de seguimiento

de SEO/BirdLife 2014, pp. 4-9. SEO/BirdLife. Madrid.

Evans, K.L. 2004. The potential for interactions between predation and habitat

change to cause population declines of farmland birds. Ibis 146: 1–13.

Fletcher, K. 2006. Predator control and ground-nesting waders. Game Conservancy

Trust Rev. 37: 80–83.

Fletcher, K., Aebischer, N.J., Baines, D., Foster, R. & Hoodless, A.N. 2010. Changes in

breeding success and abundance of ground-nesting moorland birds in

relation to the experimental deployment of legal predator control. Journal

of Applied Ecology 47: 263-272.

Galbraith, H. 1988. Effects of agriculture on the breeding ecology of lapwings

Vanellus vanellus. Journal of Applied Ecology 25: 487–503.

Gates, J.E. & Gysel, L.W. 1978 Avian nest dispersion and fledging outcome in field-

forest edges. Ecology 59: 871–83.

Gibbons, D.W., Amar, A., Anderson, G.Q.A. et al. 2007. The Predation of Wild Birds

in the UK: a Review of its Conservation Impact and Management. RSPB

Research Report No. 23. Sandy: RSPB.

188

de Graaf, R.M. & Maier, T.J. 1996. Effects of egg-size on predation by White-footed

Mice. Wilson Bulletin 108: 535–539.

Green, R.E., Tyler, G.A. & Bowden, C.G.R. 2000. Habitat selection, ranging behaviour

and diet of the stone curlew Burhinus oedicnemus in southern England.

Journal of Zoology (Lond.) 250: 161–183.

Gregory R.D., Van Strien A.J., Voriek P., Gmelig Meyling A.W., Noble D.G., Foppen

R.P.B., & Gibbins D.W. 2005. Developing indicators for European birds.

Philosophical Transactions of the Royal Society of London 360: 269-288.

Herkert J. R. 2009. Response of bird populations to farmland set-aside programs.

Conservation Biology 23: 1036–1040.

Hole, D.G., Perkins, A.J., Wilson, J.D., Alexander, I.H., Grice, P.V. & Evans, A.D. 2005.

Does organic farming benefit biodiversity? Biological Conservation 122,

113–130

Ibáñez-Álamo, J.D., Sanllorente, O. & Soler, M. 2012. The impact of researcher

disturbance on nest predation rates: a meta-analysis. Ibis 154: 5-14.

King, D.I. & DeGraaf, R.M. 2006. Predators at bird nests in a northern hardwood

forest in New Hampshire. Journal of Field Ornithology 77: 239–243.

Kleijn D., & Sutherland W.J. 2003. How effective are European agrienvironment

schemes in conserving and promoting biodiversity? Journal of Applied

Ecology 40: 947-969.

Kragten, S. & De Snoo, G.R. 2007. Nest success of Lapwings Vanellus vanellus on

organic and conventional arable farms in the Netherlands. Ibis 149: 742–

749.

Kragten, S., Trimbos, K.B. & De Snoo, G.R. 2008. Breeding skylarks (Alauda

arvensis) on organic and conventional arable farms in The Netherlands.

Agriculture, Ecosystems and Environment 126: 163–167.

189

Lapiedra O., Ponjoan A., Gamero A., Bota G., & Mañosa S. 2011. Brood ranging

behaviour and breeding success of the threatened little bustard in an

intensified cereal farmland area. Biological Conservation 144: 2882–2890.

Magaña, M., Alonso, J.C., Martín, C.A., Bautista, L.M. & Martín, B. 2010. Nest-site

selection by Great Bustards Otis tarda suggests a trade-off between

concealment and visibility. Ibis 152: 77-89.

Maier, T.J. & de Graaf, R.M. 2001. Differences in depredation by small predators

limit the use of plasticine and zebra finch eggs in artificial-nest studies.

Condor 103: 180–183.

Major, R.E., & Kendal, C.E. 1996. The contribution of artificial nest experiments to

understanding avian reproductive success: a review of methods and

conclusions. Ibis 138:298–307.

Newton, I. 1998. Population Limitation in Birds. London: Academic Press Ltd.

Newton I. 2004 The recent declines of farmland bird populations in Britain: an

appraisal of causal factors and conservation actions. Ibis 146: 579-600.

O’Connor, R.J., & Shrubb, M. 1986. Farming and Birds, Cambridge University Press.

Noske, R.A., Fischer, S. & Brook, B.W. 2008. Artificial nest predation rates vary

among habitats in the Australian monsoon tropics. Ecological Research 23:

519-527.

Opermanis, O. 2001. Marsh Harrier Circus aeruginosus predation on artificial duck

nests: a field experiment. Ornis Fennica 78: 198–203.

Paradis, E., Baillie, S.R., Sutherland, W.J., Dudley, C., Crick, H.Q.P. & Gregory, R.D.

2000. Spatial synchrony in populations of birds: effects of habitat,

population trend and spatial scale. Ecology 81: 2112–2125.

Pescador, M. & Peris, S. 2001. Effects of land-use on nest predation: an

experimental study in Spanish croplands. Folia Zoologica 50: 127–136.

190

Pita, R., Mira, A., Moreira, F., Morgado, R., & Beja, P. 2009 Influence of landscape

characteristics on carnivore diversity and abundance in Mediterranean

farmland. Agriculture, Ecosystems & Environment 132: 57-65.

Ponce, C., Bravo, C., & Alonso, J.C. 2014. Effects of agri-environmental schemes on

farmland birds: do food availability measurements improve patterns

obtained from simple habitat models?. Ecology and Evolution 4 (14): 2834-

2847.

Purger, J.J., Csuka, S. & Kurucz, K. 2008. Predation survival of ground nesting birds

in grass and wheat fields: Experiment with plasticine eggs and artificial

nests. Polish Journal of Ecology 56: 481–486.

R Development Core Team. 2013. R: A language and environment for statistical

computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-

900051-07-0, URL http://www.R-project.org.

Rangen, S.A., Clark, R.G., & Hobson, K.A. 1999. Influence of nest-site vegetation and

predator community on the success of artificial songbird nests. Canadian

Journal of Zoology 77: 1676-1681.

Reino, L., Porto, M., Morgado, R., Carvalho, F., Mira, A. & Beja, P. 2010. Does

afforestation increase bird nest predation risk in surrounding farmland?

Forest Ecology and management 260: 1359-1366.

Richardson, T.W., Gardali, T. & Jenkins, S.H. 2009. Review and meta-analysis of

camera effects on avian nest success. Journal of Wildlife Management 73:

287–293.

Schmidt, K.A. 1999. Foraging theory as a conceptual framework for studying nest

predation. Oikos 85: 151–160.

Storch, I., Woitke, E. & Krieger, S. 2005. Landscape-scale edge effect in predation

risk in forest-farmland mosaics of central Europe. Landscape Ecology 20:

927–940.

Suárez, F., Yanes, M. & Herranz, J. 1993. Nature reserves and the conservation of

iberian

191

Shrubsteppe passerines: the paradox of Nest predation. Biological Conservation

64: 77-81.

Tapper, S.C., Potts, G.R. & Brockless, M.H. 1996. The effect of an experimental

reduction in predation pressure on the breeding success and population

density of grey partridges Perdix perdix. Journal of Applied Ecology 33: 965–

978.

Tews, J., Bert, D.G. & Mineau, P. 2013. Estimated mortality of selected migratory

bird species from mowing and other mechanical operations in Canadian

agriculture. Avian Conservation and Ecology 8: 8.

Verboven, N., Ens, B.J. & Dechesne, S. 2001. Effect of investigator disturbance on

nest attendance and egg predation in Eurasian Oystercatchers. Auk 118:

503–508.

Villarías, L. 1979. Atlas de malas hierbas. Ediciones Mundiprensa, Madrid.

Whittingham, M.J., Percival, S.M. & Brown, A.F. 2002. Nestsite selection by Golden

Plover: why do shorebirds avoid nesting on slopes? Journal of Avian Biology

33: 184–190.

Whittingham, M.J. & Evans, K.L. 2004. A review of the effects of habitat structure on

predation risk of birds in agricultural landscapes. Ibis 146 (Suppl. 2): 210–

220.

Wilson, J.D. & Browne, S.J. 1993. Habitat selection and breeding success of Skylarks

Alauda arvensis on organic and conventional farmland. British Trust for

Ornithology, Thetford, UK.

Wilson, G., Harris, S. & McLaren, G.W. 1997a. Changes in the British Badger

Population, 1988–1997. London: People’s Trust for Endangered Species.

Wilson, J.D., Evans, J., Browne, S.J. & King, J.R. 1997b. Territory distribution and

breeding success of skylarks Alauda arvensis on organic and intensive

farmland in southern England. Journal of Applied Ecology 34: 1462–1478.

192

Wilson, A.M., Vickery, J.A. & Browne, S.J. 2001a. Numbers and distribution of

Northern Lapwings Vanellus vanellus breeding in England and Wales in

1998. Bird Study 48: 2–17.

Willson, M.F., Morrison, J.L., Sieving, K.E., de Santo, T.L., Santisteban, L. & Diaz, I.

2001b. Patterns of predation risk and survival of bird nests in a Chilean

agricultural landscape. Conservation Biology 15: 447–456.

Wilson, J.D., Whittingham, M.J., & Bradbury, R.B. 2005. The management of crop

structure: a general approach to reversing the impacts of agricultural

intensification on birds? Ibis 147: 453-463.

Yanes, M., Suárez, F., 1995. Nest predation patterns in ground–nesting passerines

on the Iberian Peninsula. Ecography 18:423–428.

Zuur, A.F., Ieno, E.N., Walker, N.J., Saveliev, A.A., & Smith, G.M. 2009. Mixed effects

models and extensions in ecology with R. Springer, New York.

193

DISCUSIÓN GENERAL

En esta tesis doctoral se han estudiado dos aspectos fundamentales para la

conservación de los ambientes agrícolas, como son la eficacia de la señalización en

tendidos eléctricos para reducir la mortalidad directa de las aves debida a las

colisiones y los efectos del manejo del hábitat para revertir el proceso de

intensificación agrícola. El objetivo último de ambas actuaciones es el mismo:

mejorar la sostenibilidad de las zonas agrícolas y de la biodiversidad que

mantienen, para que la conservación de las estepas agrícolas y sus comunidades de

plantas y animales sean viables a largo plazo. Para conseguir ese objetivo, en esta

tesis se han estudiado la intensificación agrícola y los tendidos eléctricos, que junto

a otras infraestructuras asociadas al desarrollo humano (carreteras, edificios, etc.),

la presión cinegética y otras causas locales, son responsables del declive de las

aves esteparias (Morales et al 2005).

Debido a la desaparición acelerada de la fauna esteparia, los planes de

recuperación y conservación de fauna y flora son aplicados en zonas remotas con

poco desarrollo (por ejemplo: Lagunas de Villafáfila, Zamora) en los que existe

fauna y flora de interés para la conservación, pero donde los impactos

antropogénicos son escasos. Para estudiar el impacto de la intensificación agrícola

y del desarrollo de los tendidos eléctricos ha sido preciso seleccionar una zona en

las afueras de una ciudad como Madrid (con más de3 millones de personas), en la

que la intensificación agrícola se pueda revertir o atenuar mediante la aplicación

de un Plan de Medidas Agroambientales. La investigación presentada en esta tesis

es, por tanto, muy infrecuente, tanto por la escala de los experimentos

(señalización de tendidos, plan de medidas agroambientales, años de trabajo)

como por desarrollarse en una zona peri-urbana, en la que la densidad de tendidos

eléctricos es propia de una gran urbe, a pesar de lo cual, sigue manteniendo una

rica comunidad de aves esteparias (incluyendo una importante población de

Avutarda común Otis tarda, Martín 2008). Las zonas degradadas

medioambientalmente en los alrededores de las grandes ciudades son frecuentes,

pero investigar en ellas la eficacia de las medidas agroambientales o el impacto de

los tendidos eléctricos suele ser difícil, debido a que no tienen la suficiente riqueza

de especies y/o sus poblaciones son inviables por su pequeño tamaño. Los

194

tendidos eléctricos son comunes en estos ambientes de la zona centro peninsular,

tanto en la Comunidad de Madrid como en la provincia de Guadalajara. No existe

ningún lugar habitado por la avifauna esteparia en estas zonas en la que falte un

tendido eléctrico, aunque la densidad de líneas eléctricas, lógicamente, varía entre

zonas (Martín 2008). En esta tesis, la zona de estudio y las especies que la habitan

se encuentran entre las que mayor cantidad de individuos contienen de todo el

rango de distribución en España y otros países de Europa (en el caso de la

avutarda). Aunque el resultado general de las medidas adoptadas puede ser

positivo tras haberse estudiado su efectividad en esta tesis, ambas cuestiones

continúan sin tener una solución plenamente satisfactoria.

Evaluación de la mortalidad de los tendidos y la eficacia de las medidas

compensatorias para reducirla

La interacción entre las aves y los tendidos eléctricos y la forma de mitigar la

mortalidad asociada a la colisión han sido muy estudiadas durante los últimos 50

años (ver revisión en Barrientos et al. 2011), a pesar de lo cual no se han obtenido

resultados plenamente satisfactorios en ningún trabajo, en cuanto a la eliminación

total de este factor de mortalidad. Este hecho pone de manifiesto la complejidad

del problema que suponen estas infraestructuras para las aves. A esto se debe

añadir la falta de estandarización de los métodos de recogida de la información y la

falta de robustez en algunos de los análisis efectuados (Bevanger 1999). Además,

existen varios sesgos que pocas veces se han considerado en este tipo de estudios,

por lo que sus conclusiones y extrapolaciones a otras zonas se deben tomar con

cierta cautela.

El trabajo mostrado en esta tesis permitió desarrollar una metodología para

calcular de qué manera se subestima la cantidad real de aves muertas en tendidos

eléctricos y los factores implicados (capítulo 1), así como estimar la mortalidad

producida en las líneas eléctricas y estudiar la eficacia de la colocación de

dispositivos anticolisión (capítulo 2).

Los estudios sobre mortalidad en tendidos eléctricos tienen como primer

objetivo conocer la magnitud del problema que generan la colisión y/o la

electrocución en una zona determinada. Por otro lado, cuando se señaliza un

195

tendido eléctrico el objetivo es reducir la mortalidad de aves. Según se ha puesto

de manifiesto en esta tesis, es fundamental llevar a cabo varios tipos de

experimentos (desaparición de cadáveres y detectabilidad según el observador),

puesto que existen sesgos que hacen variar en gran medida los resultados

obtenidos.

En nuestro caso, encontramos diferencias en la desaparición de cadáveres

según el tamaño de los ejemplares. Aunque se ha propuesto una ecuación general

válida para los diferentes tamaños (Figura 4, capítulo 2), es probable que existan

grandes variaciones en la tasa de desaparición respecto a otros hábitats. Aunque el

objetivo principal de ese capítulo no incluía este aspecto, sí debe considerarse a la

hora de extrapolar las ecuaciones obtenidas a otros lugares.

En estudios científicos y técnicos sobre interacciones entre aves e

infraestructuras suelen participar personas con muy diversa formación en la

búsqueda de cadáveres. Tal y como muestran los resultados de esta tesis, sería

recomendable cambiar este aspecto, ya que los resultados obtenidos podrían estar

sesgados, aumentando el número de ejemplares localizados por observadores más

experimentados (Figura 5, capítulo 2). En este caso, los resultados obtenidos no

serían válidos si no se corrigen por la experiencia en la localización e identificación

de los restos. Por tanto, existen tres opciones para evitar este sesgo. Una es que los

participantes en este tipo de proyectos tengan la misma formación de partida. Otra

es realizar un número alto de sesiones en búsqueda de cadáveres bajo los tendidos

eléctricos previos al comienzo de los muestreos para el estudio. Por último, otra

posibilidad es que se lleven a cabo experimentos similares a los de esta tesis

doctoral y se apliquen las correspondientes correcciones. Sin duda, este último

caso aportaría un valor extra a cualquier trabajo de este tipo.

La investigación llevada a cabo para conocer la eficacia de la señalización en

tendidos eléctricos (capítulo 2) ha permitido aplicar los valores de desaparición de

cadáveres del capítulo 1. Los resultados obtenidos mediante el número de

cadáveres encontrados fueron muy diferentes de los obtenidos mediante las

estimas aplicando las correcciones necesarias (Tabla 4, capítulo 2).

196

Pudimos estimar de una forma precisa cuántas aves mueren en los tendidos

eléctricos de zonas agrícolas del centro peninsular. Además, se obtuvo un listado

de especies realmente amplio (Tabla 3, capítulo 2), algunas de ellas protegidas,

como la avutarda, Sisón común (Tetrax tetrax), Buitre negro (Aegypius monachus)

o Aguilucho lagunero occidental (Circus aeruginosus). La señalización fue eficaz,

incluso cuando se desciende a nivel de especie. Por ejemplo, ésta fue más efectiva

para el sisón y, algo menos para la avutarda. Es necesario recordar que ambas se

encuentran amenazadas según diferentes catálogos, tanto a nivel regional como

nacional, son especies representativas de los ambientes agrícolas del centro

peninsular y parecen proclives a las colisiones con tendidos eléctricos (Alonso et

al. 1994, Barrientos et al. 2011, Ferrer 2012). Por tanto, estudiar la eficacia de la

señalización ha permitido calcular la magnitud real sobre estas y otras especies

que viven en medios agrícolas.

Aunque la señalización ha sido eficaz a la hora de disminuir la mortalidad

de las aves, no se ha eliminado completamente este factor de mortalidad. Es

necesario estudiar cuál es la influencia de los tendidos eléctricos en la viabilidad

poblacional de especies amenazadas (Bevanger 1999) antes (Martín 2008) y

después de la señalización. En este sentido cabe decir que el equipo de

investigación en el que se ha desarrollado esta tesis doctoral dispone de una serie

larga de años de censos para la avutarda, además de numerosos datos sobre

ejemplares marcados muertos en los tendidos eléctricos de toda España. Estas

bases de datos tienen un valor creciente y permitirán en el futuro llevar a cabo

esos análisis post-señalización, así como otros sobre la localización de lugares

concretos donde la mortalidad es más elevada dentro de un tramo de tendido

(puntos de riesgo máximo o puntos negros). De esta manera, si se produjera una

nueva señalización o refuerzo de la ya existente, se podrían dirigir los esfuerzos

hacia esos lugares, sin tener que emplear altos presupuestos económicos.

El 29 de Agosto de 2008 se aprobó el real Decreto 1432/2008, por el que se

establecen medidas para la protección de la avifauna contra la colisión y la

electrocución en líneas eléctricas de alta tensión. Dicho decreto establece la

obligatoriedad de marcaje o modificación de tendidos eléctricos peligrosos en

zonas protegidas. Sin embargo, la obligación se refiere exclusivamente a la

197

electrocución de aves, sin considerar la colisión, cuya mitigación será voluntaria.

Tal y como se ha puesto de manifiesto en esta tesis y otras investigaciones (Ferrer

2012), la colisión es un grave problema para varias especies de aves esteparias. No

se entiende que en el real Decreto no se haya incluido la colisión de aves como una

amenaza de similar magnitud a la producida por la electrocución en líneas

eléctricas de zonas protegidas (Íñigo et al. 2010). Sería necesario hacer una

revisión del Real Decreto para subsanar este grave error.

Además, el real decreto deja fuera del ámbito de aplicación de las medidas

todas las zonas que no estén incluidas en zonas de protección. En la Comunidad de

Madrid existen varios lugares que quedan fuera de esas zonas protegidas, pero que

albergan importantes grupos de avutardas y sisones (más susceptibles que otras

especies a la colisión), como son Campo Real, Fuentidueña-Estremera de Tajo, o

Torrejón de Velasco.

Evaluación de la eficacia de las medidas agroambientales para reducir los

efectos de la intensificación agrícola

Las zonas agroesteparias suponen una gran parte de la superficie europea y

española, por lo que su gestión afecta a la conservación de multitud de organismos,

tanto plantas como animales. Por otro lado, los cambios producidos en la

agricultura durante las últimas décadas han desembocado en severos problemas

ambientales para distintos grupos de seres vivos, hasta el punto en que se

considera a la intensificación la principal causante de dichos problemas. Para

contrarrestar ese efecto negativo se ha actuado mediante Programas de Medidas

Agroambientales a nivel continental y estatal (Carricondo et al. 2012), aunque los

resultados varían enormemente entre regiones (Kleijn & Sutherland 2003).

Los resultados proporcionados en esta tesis apoyan la idea general de que

el manejo del hábitat a través de medidas agroambientales puede llegar a aportar

beneficios ambientales a los sistemas agrícolas (Kleijn et al. 2006). Se han

implementado medidas para favorecer el tamaño y composición de las especies de

plantas, artrópodos (capítulo 3) y aves (capítulos 4 y 5). Sin embargo, la eficacia de

cada una de las medidas adoptadas no es igual en cada grupo de organismos. En la

presente tesis se pone de manifiesto la necesidad de reducir la intensificación

198

agrícola para favorecer a los organismos de los sistemas cerealistas de secano

mediterráneos (Sans et al. 2013). La cuestión más complicada es, seguramente,

decidir de qué manera se debe llevar a cabo ese proceso para revertir la situación

actual sin perjudicar la producción agrícola, tendente al incremento (Wilson et al.

2009, Pretty et al. 2010). Existen numerosas medidas que, además, varían en la

forma de implantación en cada uno de los países (Kleijn et al. 2006) e incluso a

escala regional dentro de cada uno. Basta ver las medidas desarrolladas por la

Comunidades Autónomas (Carricondo et al. 2012).

En el capítulo 3 de la tesis se lleva a cabo un análisis de la eficacia del

manejo en parcelas de cultivo de cereal. Concretamente se estudia la respuesta de

plantas silvestres y artrópodos al cultivo ecológico de cereal en comparación con el

manejo convencional. El manejo ecológico del cereal fue positivo para ambos

grupos (Tablas 3 y 5), a pesar de que el manejo ecológico sólo se ha producido

durante un año, cuya consecuencia es que los herbicidas empleados en años

anteriores persisten. Se incrementaron todos los valores analizados respecto a las

plantas silvestres (Tabla 3) y al comparar los resultados con los de otras regiones

se obtiene que el incremento en parámetros de vegetación es mucho mayor en

nuestra zona de estudio. La Región Mediterránea es más rica en plantas silvestres,

la agricultura de nuestra zona no es tan intensiva como en otros lugares y las

parcelas son de pequeño tamaño. Todo ello hace que se den las condiciones

apropiadas para una mayor proliferación de plantas silvestres. También

aumentaron la mayor parte de los parámetros relacionados con los artrópodos

(Tabla 4). Tanto la abundancia, como la riqueza y la biomasa (esta ligeramente).

Lógicamente, esto se relaciona con la eliminación del uso de biocidas en las

parcelas cultivadas de forma ecológica. Sin embargo, una consecuencia del manejo

ecológico fue la proliferación de varios grupos de artrópodos, lo que hizo que la

diversidad en siembras convencionales fuera mayor que en la siembras cultivadas

de forma ecológica.

El aumento de esos parámetros mediante el cultivo ecológico de cereal ha

producido consecuencias adversas en la depredación de nidos artificiales (Figura

2, capítulo 5). Los nidos localizados en parcelas cultivadas de forma ecológica y

aquellos que estaban rodeados por mayor superficie de cultivo de cereal ecológico

199

vieron incrementada su tasa de depredación (Tabla 2). En cambio, las siembras de

cereal convencional no produjeron efectos positivos en la mayoría de los

parámetros estudiados sobre plantas, artrópodos y aves (capítulo 3, capítulo 4),

pero sufrieron las menores tasas de depredación de nidos (Figura 2, capítulo 5).

Así pues, se genera un conflicto entre los beneficios ambientales y la depredación

de nidos. Tal y como han mostrado otros autores (Redisma et al. 2006, Armengot

et al. 2011), existe un gradiente en el manejo del cultivo de cereal, tanto

convencional como ecológico. Para compaginar ambas cuestiones, y obtener

resultados positivos en ambos casos, sería necesaria la adopción de una solución

de compromiso. Es decir, reducir el manejo intensivo en las siembras de cereal

convencional y no llegar al extremo del cultivo ecológico. De esta forma sería

esperable un aumento del valor ecológico de las siembras convencionales, aunque

también un incremento en la tasa de depredación, lo contrario para las siembras

manejadas de una forma más extensiva. Esta hipótesis no ha sido comprobada en

esta tesis, pero los resultados obtenidos permiten su formulación para comprobar

su veracidad en futuros experimentos.

El análisis de los parámetros durante todo el ciclo anual de las aves reflejó

que el manejo de parcelas incluidas en el programa de medidas agroambientales

resultó positivo para todas las épocas del año (Tabla 2, capítulo 4). Sin embargo,

otras variables no incluidas en el programa también resultaron importantes,

probablemente debido a que la zona de estudio no está especialmente

intensificada (Concepción et al. 2008, Concepción et al. 2012). Por tanto, uno de los

primeros resultados destacables de este capítulo es la necesidad de considerar

variables a una escala mayor que las propias medidas agroambientales

implantadas.

Los dos grupos de modelos retuvieron básicamente similares predictores

con independencia de la dificultad de medir las variables que los integraban

(modelos complejos vs modelos sencillos). Sin embargo, los modelos complejos,

que incluían variables costosas de medir (Tabla 5), tuvieron en general mejores

resultados a la hora de predecir la respuesta de las aves ante el manejo agrícola

(Tabla 4) que los modelos sencillos. Los modelos complejos permitieron entender

de una manera más precisa los factores que subyacen a esas respuestas. Hay varios

200

casos en los que estos modelos revelaron la importancia de las variables de

comida, más que la estructura o la superficie de la parcela, que no habrían salido a

la luz si se hubiera empleado el conjunto de modelos de hábitat (modelos

sencillos). Sin embargo, el esfuerzo y dinero necesarios para llevar a cabo los

modelos complejos son altos en comparación con los modelos sencillos de hábitat.

Sugerimos que el desarrollo de los programas de medidas agroambientales que se

lleven a cabo en el futuro consideren la posibilidad de incluir un presupuesto

específico para poder registrar en campo las variables necesarias para el cálculo de

los modelos complejos, de manera que sea posible una evaluación científica más

precisa de la eficacia de dichos programas.

El trabajo de campo desarrollado ha puesto de manifiesto la importancia de

semillas y artrópodos para la dieta de las especies de aves esteparias durante el

invierno. Aunque este hecho ya ha sido propuesto en otros estudios (Evans et al.

2011), es necesario destacar que las semillas empleadas en el programa de

medidas agroambientales de esta tesis doctoral no estuvieron sometidas a ningún

tratamiento fitosanitario, gracias a lo cual no se observaron los daños a las aves

descritos en otros estudios (López-Antia et al. 2015).

La medida más efectiva para las variables medidas en el capítulo 4 fue el

barbecho de larga duración (es decir, la retirada de la producción de una parcela

agrícola) debido a la gran cantidad de alimento que aporta mediante semillas,

artrópodos y plantas (Chamberlain et al.1999, Henderson et al. 2000, Lapiedra et

al. 2011). Sin embargo, también fue uno de los tipos de parcela más depredados,

casi tanto como las parcelas labradas (Figura 2, capítulo 5). El hecho de que

parcelas de alta calidad (capítulo 3 y 4) sean notablemente más depredadas que las

siembras convencionales sugiere la posibilidad de la presencia de trampas

ecológicas (Donald 1999, Evans 2004), debido a un fenómeno de atracción de

depredadores a los lugares con mayor densidad de comida (aves, micromamíferos,

nidos, etc.). Los planes de conservación de aves ligadas a estos medios deben

considerar la interacción de ambos efectos a la hora de implementar los barbechos

de larga duración. Según los resultados obtenidos, las parcelas de barbecho y de

cultivo ecológico de cereal deberían dispersarse en grandes áreas. De esta forma se

mantendrían los efectos positivos de las medidas y se reduciría, probablemente, la

201

depredación (Bayne et al. 1997). Existe literatura sobre el impacto que tiene la

concentración de parcelas de alto valor ecológico en una matriz agrícola (por

ejemplo, bosques isla en las comunidades de aves, Santos et al 2002), pero en el

caso de las aves esteparias se desconoce cómo influye la dispersión de parcelas

valiosas en su comportamiento y la evolución de sus poblaciones.

El cultivo de leguminosa también fue importante, tanto la biomasa de

semillas de veza y los artrópodos como la superficie cultivada presentes en todas

las épocas del año y para muchas de las variables respuesta consideradas (Tabla 3,

capítulo 4). Es de sobra conocido que las leguminosas son una fuente importante

de alimento para las aves esteparias (Bretagnolle et al. 2011, Bravo et al. 2012). Lo

que no se había estudiado hasta la fecha es que la tasa de depredación producida

en este cultivo es la más baja respecto a cualquier otra medida agroambiental

aplicada (Figura 2, capítulo 5). Además, el beneficio de la semilla no tratada hace

de esta medidas muy útil y necesaria para las aves de zonas agrícolas.

Las parcelas labradas aparecieron también como factor importante en los

análisis de abundancia y riqueza de aves (Tabla 2, capítulo 4). Especies como la

alondra común o la avefría europea prefieren este tipo de sustrato para

alimentarse debido a que su estrategia frente a los depredadores implica tener un

gran campo visual y poder huir antes de que el depredador esté cerca (Butler et al.

2005). Sin embargo, este tipo de parcela sufrió la mayor tasa de desaparición,

relacionada con la escasez de vegetación donde poder cobijar los nidos (Figura 3,

capítulo 5). Determinadas aves, como las mencionadas anteriormente, emplean

este tipo de parcelas también durante la época de nidificación, basando su

estrategia antidepredadora en el mismo procedimiento descrito anteriormente

(Whittingham et al. 2002). A la alta tasa de depredación hay que sumarle el riesgo

de ser destruidos durante las labores agrícolas de los tractores. Una especie que

nidifique en este sustrato tiene altas probabilidades de que su huevos no puedan

llegar a eclosionar. Es necesario abordar este problema con rapidez debido a su

alto impacto sobre las aves ligadas a medios abiertos y las consecuencias directas

hacia las poblaciones. En esta tesis se proponen varias medidas que podrían

resultar eficaces. Sería necesario restringir las labores agrícolas durante la época

de nidificación, lo que evitaría la destrucción directa de nidos. Además, sería

202

recomendable favorecer un cierto grado de desarrollo vegetal el cual ayudaría a

que las aves nidificantes en este sustrato dispongan de lugares más protegidos y

mayor alimento. Por último, en algunas zonas existe un ciclo agrícola demasiado

acoplado. Un año la mayor parte de las parcelas están cultivadas, y al siguiente

todas ellas son parcelas labradas. Si se desacoplara el ciclo agrícola es posible que

ello redundase en menores tasas de depredación y en un aumento importante de la

cantidad de alimento disponible para las aves, no sólo a escala de parcela, sino

también de paisaje.

La colocación de los nidos en el borde y el interior de la parcela evidenció

una depredación diferencial (Tabla 2, capítulo 5). Los nidos del borde de la parcela

tuvieron mayor riesgo de ser depredados. Este hecho, junto con los resultados de

la estructura de la vegetación (menor depredación cuanto más alta es la

vegetación), y la determinación de los grupos de depredadores, ayudaron a

explicar los patrones obtenidos. En nuestro estudio, la mayor parte de los eventos

de depredación los protagonizaron las aves. Éstas no depredaron más sobre un

tipo de parcela concreta, ni se detectaron diferencias respecto a la localización del

nido dentro de la parcela. El motivo es que la estrategia de búsqueda de alimento

de las aves (sobre todo las rapaces registradas) se basa en prospectar el territorio

en vuelo y detectar a sus presas de forma visual. Los mamíferos, en cambio, sí

depredaron más sobre los barbechos y en nidos localizados en el borde de la

parcela. Esto se debe a que los mamíferos se basan en el olfato para detectar a sus

presas y, además, utilizan las estructuras lineales para moverse entre zonas. El

control de depredadores no es recomendable en nuestra zona de estudio. A pesar

de no tener información sobre este parámetro, la ausencia de zorros (Vulpes

vulpes) nos indica que ya existe un control de depredadores en la zona. Por otro

lado, los aguiluchos fueron los depredadores que más aparecieron en las cámaras.

Ambas especies de aguiluchos (cenizo y lagunero) están estrictamente protegidas.

El aguilucho cenizo está amenazado y se están llevando a cabo medidas para

favorecerlo, como el salvamento de nidos o el retraso en la recogida de cereal.

Sugerimos que se haga un manejo del hábitat para dificultar el acceso de los

depredadores a los nidos mediante el desarrollo de la vegetación y, por tanto, la

mayor ocultación de los nidos (Davis 2005).

203

REFERENCIAS

Alonso JC, Alonso JA, Muñoz-Pulido R (1994) Mitigation of bird collisions with

transmission lines through groundwire marking. Biological Conservation

67: 129-134.

Armengot, L., José-María, L., Blanco-Moreno, J.M., Bassa, M., Chamorro, L. & Sans,

F.X. 2011. A novel index of land use intensity for organic and conventional

farming of Mediterranean cereal fields. Agronomy for Sustainable

Development 31: 699-707.

Barrientos, R., Alonso, J.C., Ponce, C. & Palacín, C. 2011. Meta-analysis of the

effectiveness of marked wire in reducing avian collisions with power lines.

Conservation Biology 25: 893-903.

Bayne, E.M., Hobson, K.E. & Fargey, P. Predation on artificial nests in relation to

forest type: contrasting the use of quail and plasticine egges. Ecography 20:

233-239

Bevanger, K. 1999. Estimating bird mortality caused by collision and electrocution

with power lines; a review of methodology. pp. 29-56. En Birds and power

lines: collision, electrocution, and breeding. Quercus, Madrid, Spain.

Bravo, C., Ponce, C. Palacín, C. & Alonso, J.C. 2012. Diet of young Great Bustards Otis

tarda in Spain: sexual and seasonal differences. Bird Study 59:243–251.

Bretagnolle, V., Villers, A., Denonfoux, L., Cornulier, T., Inchausti, P. &

Badenhausser, I. 2011. Rapid recovery of a depleted population of Little

Bustards Tetrax tetrax following provision of alfalfa through an agri-

environment scheme. Ibis 153:4–13.

Carricondo, A., Martínez, P. & Cortés, Y. 2012. Evaluación global de las medidas

agroambientales para aves esteparias en España (2007-2013): Proyecto

Ganga (informe completo). SEO/BirdLife. Madrid.

204

Chamberlain, D.E., Wilson, A.M., Browne, S.J. & Vickery, J.A. 1999. Effects of habitat

type and management on the abundance of skylarks in the breeding season.

Journal of Applied Ecology 36: 856-870.

Concepcón, E.D., Díaz, M. & Baquero, R.A. 2008. Effects of landscape complexity on

the ecological effectiveness of agri-environment schemes. Landscape

Ecology 23: 135-148.

Concepcón, E.D., Díaz, M, Kleijn, et al. 2012. Interactive effects of landscape context

constrain the effectiveness of local agri-environmental management.

Journal of Applied Ecology 49: 695-705.

Davis, S.K. 2005. Nest-site selection patterns and the influence of vegetation on

nest survival of mixed-grass prairie passerines. Condor 107: 605-616.

Donald, P.F. 1999. The Ecology and Conservation of Skylarks Alauda arvensis on

Lowland Farmland. DPhil thesis, Oxford University.

Evans, K.L. 2004. The potential for interactions between predation and habitat

change to cause population declines of farmland birds. Ibis 146: 1-13.

Evans, D. M., Pocock, J.O., Brooks, J. & Memmott, J. 2011. Seeds in farmland food-

webs: resource importance, distribution and the impacts of farm

management. Biological Conservation 144: 2941-2950.

Ferrer, M. 2012. Aves y tendidos eléctricos. Del conflicto a la solución. Endesa S.A.y

Fundación Migres, Sevilla, 2012.

Henderson, I.G., Cooper, J., Fuller, R.J. & Vickery, J.A. 2000. The relative abundance

of birds on set-aside and neighbouring fields in summer. Journal of Applied

Ecology 37: 335-347.

Íñigo, A., Infante, O., López, V., Valls, J. & Atienza, J.C. 2010. Directrices para la

redacción de Planes de Gestión de la Red Natura 2000 y medidas especiales

a llevar a cabo en las ZEPA. SEO/BirdLife, Madrid.

205

Kleijn, D., and W. J. Sutherland. 2003. How effective are European agri-

environment schemes in conserving and promoting biodiversity? Journal of

Applied Ecology. 40: 947-969.

Kleijn, D., Baquero, R.A., Clough, Y. et al. 2006. Mixed biodiversity benefits of agri-

environment schemes in five European countries. Ecology Letters 9: 243-

254.

Lapiedra, O., Ponjoan, A., Gamero, A., Bota, G. & Mañosa, S. 2011. Brood ranging

behaviour and breeding success of the threatened little bustard in an

intensified cereal farmland area. Biological Conservation 144: 2882-2890.

López-Antia, A., Ortiz-Santaliestra, M.E., Mougeot, F. & Mateo, R. Imidacloprid-

treated seed ingestion has lethal effect on adult partridges and reduces both

breeding investment and offspring immunity. Environmental Research 136:

97-107

Martín, B. 2008. Dinámica de población y viabilidad de la Avutarda común en la

Comunidad de Madrid. Tesis doctoral.

Morales, M.B., Mañosa, S., Camprodón, J. & Bota, G. 2005. Ecology and Conservation

of Steppe-land Birds . International Symposium on ecology and

conservation of steppe-land birds. Lleida, Spain. Lynx edicions. 343 págs.

Palacín, C., Martín, C.A., Martín, B., Ponce, C. & Sastre, P. 2006. Abundancia y

distribución invernal de la Ganga Ortega (Perocles orientalis) y la Ganga

Ibérica (Pterocles alchata) en la Comunidad de Madrid. In: De la Puente, J.,

Pérez-Tris, J. & Bermejo, A. (eds) Anuario Ornitológico de Madrid 2005.

SEO-Monticola, Madrid, pp 68-75.

Pretty, J., Sutherland, W.J. Ashby. J. et al. 2010. The top 100 questions of

importance to the future of global agriculture. International Journal os

Agricultural Sustainability 8: 219-236.

REAL DECRETO 1432/2008, de 12 Septiembre. Ministerio de la Presidencia. BOE

222, de 13 de Septiembre.

206

Reidsma, P., Tekelenburg, T., van den Berg, M. & Alkemade, R. 2006. Impacts of

land-use change on biodiversity: An assessment of agricultural biodiversity

in the european Union. Agriculture, Ecosystems and Environment 114: 86-

102.

REGLAMENTO (UE) nº 1305/2013 DEL PARLAMENTO EUROPEO Y DEL CONSEJO

de 17 de diciembre de 2013 relativo a la ayuda al desarrollo rural a través

del Fondo Europeo Agrícola de Desarrollo Rural (Feader) y por el que se

deroga el Reglamento (CE) no 1698/2005 del Consejo.

Sans, F.X., Armengot, L., Bassa, M., Blanco-Moreno, J.M., Caballero-López, B.,

Chamorro, L. & José-María-L. 2013. La intensificación agrícola y la

diversidad vegetal en los sistemas cerealistas de secano mediterráneos:

implicaciones para la conservación. Ecosistemas 22(1): 30-35.

Santos, T., Telleria, J.L., Carbonell, R. 2002. Bird conservation in fragmented

Mediterranean forests of Spain: effects of geographical location, habitat and

landscape degradation. Biological Conservation 105: 113-125.

Whittingham, M.J., Percival, S.M. & Brown, A.F. 2002. Nestsite selection by Golden

Plover: why do shorebirds avoid nesting on slopes? Journal of Avian Biology

33: 184–190.

Wilson, J.D., Evans, A.D., Grice, P.V. 2009, Bird Conservation and Agriculture,

Cambridge University Press, Cambridge, UK.

207

CONCLUSIONES

11 Los estudios sobre mortalidad de aves en tendidos eléctricos llevan asociados

varios sesgos que infravaloran la cantidad real de aves muertas. Se desarrollan

y proponen varios índices de corrección para la desaparición de cadáveres y

para su detectabilidad en función del tamaño de las aves y de la experiencia de

los observadores, y se plantean las frecuencias de revisión de tendidos más

adecuadas.

22 La mortalidad de aves colisionadas contra tendidos eléctricos supone un grave

problema en las estepas cerealistas del centro peninsular. La señalización de

tendidos eléctricos reduce la mortalidad de forma significativa, aunque no la

elimina, desconociéndose además las implicaciones sobre la viabilidad

poblacional en de especies amenazadas.

33 El manejo del hábitat agrícola mediante medidas agroambientales beneficia a

plantas silvestres, artrópodos y aves. Sin embargo se debe profundizar en el

efecto de las medidas implementadas sobre las especies amenazadas.

44 El cultivo de cereal ecológico beneficia a plantas y artrópodos, aunque produce

un incremento en la tasa de depredación de nidos. La reducción de la

intensificación en siembras convencionales podría proporcionar mayor

cantidad de alimento, así como otros beneficios, a los grupos considerados. Sin

embargo, es probable que repercuta de manera negativa sobre la tasa de

depredación de nidos.

55 Las aves esteparias se ven favorecidas durante todo el ciclo anual por las

medidas agroambientales. Sin embargo, otras variables relacionadas con el

paisaje agrícola también son influyentes. El desarrollo de modelos complejos,

aunque costosos, permite conocer los factores subyacentes a la respuesta de

las aves de una forma más precisa que modelos más sencillos. Así, las aves

esteparias se ven favorecidas más por la cantidad de alimento presente que

por la composición del mosaico de cultivos.

208

66 La depredación de nidos está influida por variables a escalas de paisaje,

parcela y lugar de nidificación. Las medidas agroambientales no logran reducir

la tasa de depredación de nidos, debido a que las medidas atraen a los

depredadores. Para reducir la tasa de depredación se debe incrementar la

altura de la vegetación y distribuir las parcelas con medidas de forma más

dispersa.

77 España tiene un mayor compromiso que otros países de Europa en la

conservación de los hábitats agrícolas y su biodiversidad asociada. Alberga

algunas especies ausentes o escasas en otros países y el manejo del hábitat en

nuestras latitudes tiene consecuencias en la conservación a escala continental.

209

ANEXO 1

Programa de Medidas Agroambientales del Área Importante para las Aves

Talamanca-Camarma.

210

211

Programa de Medidas Agroambientales

del Área Importante para las Aves

Talamanca-Camarma

Dirección de contacto:

Carlos Ponce Cabas

Programa de Medidas Agroambientales del AIA Talamanca-Camarma

Museo Nacional de Ciencias Naturales

C/ José Gutiérrez Abascal, 2 - 28006 MADRID

Tel. 91 411 13 28 (extensión 1111) Fax 91 564 50 78

carlosp@mncn.csic.es

Museo Nacional de Ciencias Naturales Consejo Superior de Investigaciones Científicas

Autopista del Henares, S.A. Concesionaria del Estado

212

INTRODUCCIÓN

Las áreas de carácter estepario ocupan una buena parte del territorio de las

provincias de Madrid y Guadalajara. La mayoría de ellas tienen su origen en las

prácticas agrícolas y ganaderas que durante milenios han trasformado los bosques

primitivos en extensos campos dedicados a cultivos de secano o pastizales.

La agricultura extensiva tradicional permitía la coexistencia de explotaciones

agrarias dedicadas al cultivo de herbáceos de secano con una rica biodiversidad, en

especial, con importantes comunidades de aves esteparias. El reciente proceso de

intensificación de la agricultura ha modificado el equilibrio existente entre hombre y

aves en las llanuras, poniendo en peligro a gran parte de las especies de aves

esteparias.

Las importantes poblaciones de diferentes especies de aves esteparias que

todavía habitan en la península Ibérica pueden considerarse únicas a escala europea,

por lo que su conservación depende en gran medida del mantenimiento de los

ecosistemas agroesteparios ibéricos.

La pérdida y fragmentación del hábitat agroestepario ocasionada por la

construcción de grandes infraestructuras es otra de las principales causas de

disminución de la superficie de ecosistemas agroesteparios. Los efectos negativos de

dichas intervenciones humanas se derivan, por una parte, de la pérdida neta de

superficie disponible para las especies, y por otra, de la fragmentación del hábitat,

que resulta dividido en unidades cada vez menores.

Con el fin de compatibilizar la actividad agrícola con la conservación de la

naturaleza, y para compensar la pérdida de hábitat ocasionada por la construcción y

explotación de las autopistas R-2 y M-50, se propone el presente Programa de

Medidas Agroambientales del Área Importante para las Aves “Talamanca-

Camarma”. Este Programa está incluido en el Proyecto de medidas preventivas, correctoras y

compensatorias de la afección de la M-50 (tramo M-607 / N-IV, subtramo N-I / N-II) y de la

Autopista de peaje R-2 a la población de avutardas y otras aves esteparias de la IBA

“Talamanca-Camarma”, y al LIC “Cuenca de los ríos Jarama y Henares”.

213

OBJETIVOS DEL PROGRAMA

El objetivo principal de este Programa es compatibilizar la conservación de las

poblaciones de aves esteparias con la explotación agrícola de secano dentro de las

provincias de Madrid y Guadalajara. A través de un sistema de primas económicas se

pretende beneficiar a los agricultores que utilicen métodos de producción agraria

compatibles con la conservación de la biodiversidad.

ÁMBITO DE APLICACIÓN

Las actuaciones se llevarán a cabo en un total de siete zonas (ver anexo 1) situadas

dentro del Área Importante para las Aves denominada “Talamanca-Camarma”, en las

provincias de Madrid y Guadalajara. Seis zonas estarán situadas en la provincia de Madrid,

prácticamente todas ellas dentro de la Zona de Especial Protección para las Aves “Estepas

cerealistas de los ríos Jarama y Henares”, y una en la provincia de Guadalajara, dentro de la

ZEPA “Estepas cerealistas de la Campiña”. Se han seleccionado estas siete zonas, por ser

las que cuentan con mayor diversidad y tamaño de poblaciones de aves esteparias y por

estar directamente afectadas por la construcción de las autopistas R-2 y M-50.

BENEFICIARIOS

Podrán acogerse voluntariamente al Programa de ayudas todos los agricultores con

tierras de secano dedicadas al cultivo de herbáceos incluidas en el ámbito de aplicación del

Programa. En el caso de aquellos agricultores que explotan tierras en régimen de

arrendamiento o aparcería tendrán que actuar de acuerdo con el propietario.

Este Programa va dirigido a las superficies dedicadas al cultivo de herbáceos de

secano. Se excluyen cultivos de regadío, cultivos leñosos y superficie de erial, pastizal o

arbolado, así como terrenos improductivos. Sin embargo, es perfectamente compatible con

los programas de ayudas agroambientales derivados de los Reglamentos 2078/92 y

1257/1999 de la CEE y del Real Decreto 4/2001.

Los beneficiarios podrán elegir entre las cinco medidas que aparecen descritas a

continuación.

214

COMPROMISOS DE LAS MEDIDAS

Medida 1: Mejora y mantenimiento del barbecho tradicional

Para el conjunto de parcelas acogidas a esta medida, que estarán destinadas a

barbecho, el agricultor se comprometerá a:

Mantener los rastrojos sin alzar desde la cosecha de cereal precedente, en el mes de

julio, hasta el 1 de enero. A partir de esta fecha el agricultor podrá labrar los

barbechos sin aplicar productos fitosanitarios ni ninguna otra sustancia química

hasta el 31 de marzo

Nuevamente durante los meses de abril, mayo y junio no podrá realizar ninguna

labor agrícola sobre los barbechos acogidos a esta medida

Medida 2: Barbecho semillado con leguminosas

Para las parcelas acogidas a esta medida, el agricultor se comprometerá a:

Siembra de leguminosas (veza, yeros, alfalfa, guisantes, garbanzos…) sobre parcelas

destinadas a barbecho

Preparar el terreno correctamente para el buen desarrollo de las plantas de

leguminosa

Comunicar la fecha de siembra al responsable del Programa de Medidas

Agroambientales al menos 5 días antes de realizarla.

No emplear más del 20% de semilla de cereal junto con la semilla de leguminosa

No utilizar semillas tratadas o blindadas para la sementera.

La siembra debe ser realizada en el mes de octubre

Enterrado de dichos barbechos semillados no antes del 10 de julio

No utilizar abonos ni productos fitosanitarios durante el período de duración del

barbecho semillado

215

Medida 3: Retirada de la producción de tierras durante el periodo de duración del

programa (Máximo de 5 años para los acogidos en el ciclo agrícola 2006-2007)

Para las parcelas acogidas a esta medida, el agricultor se comprometerá a:

No realizar labores agrícolas durante el periodo establecido de retirada de la

producción

No utilizar productos fitosanitarios durante el período de retirada

No quemar el barbecho durante el período de retirada

Para acogerse a esta medida es necesario entregar una declaración o comprobante

del uso agrícola de la tierra durante los últimos 3 años.

Medida 4: Rotación de cultivos trigo- girasol

Para las parcelas acogidas a esta medida, el agricultor se comprometerá a:

Mantener los rastrojos procedentes de la siembra del cereal precedente, sin alzar

hasta al menos el 1 de enero

La siembra de girasol será efectuada en un intervalo de tiempo que comprende

desde el 1 de enero hasta el 31 de marzo, sin poderse prorrogar con posterioridad a

esta ultima fecha

Preparar el terreno correctamente para el buen desarrollo de las plantas de

leguminosa

Comunicar la fecha de siembra al responsable del Programa de Medidas

Agroambientales al menos 5 días antes de realizarla

La siembra se realizará en cantidades no inferiores a 3,250 Kg. por hectárea o

45.000-50.000 plantas por hectárea, con una separación entre líneas de cultivo de

aproximadamente 70 cm

Los agricultores acogidos a esta medida se comprometerán a no utilizar herbicidas

en el cultivo del girasol

Las labores de triturado y enterrado del cañote del girasol no podrán ser efectuadas

antes del 30 de septiembre

216

Medida 5: Cultivo de cereal no tratado

Los agricultores acogidos a esta medida deberán de respetar los siguientes

compromisos:

Utilizar para la sementera exclusivamente semillas que no contengan productos

fitosanitarios (semillas no blindadas ni tratadas)

Preparar el terreno correctamente para el buen desarrollo de las plantas de

leguminosa

Comunicar la fecha de siembra al responsable del Programa de Medidas

Agroambientales al menos 5 días antes de realizarla.

No realizar tratamientos fitosanitarios sobre la parcela acogida durante el periodo

de duración del compromiso (desde su siembra en el mes de octubre, hasta su

retirada a partir del 10 de julio)

La siembra se realizará en las fechas habituales para el cereal, no entrando a realizar

ningún tipo de labor ni ninguna práctica, que contribuya a espantar a la fauna de las

parcelas acogidas hasta el final del compromiso que será el 10 de julio.

Los agricultores acogidos a esta medida estarán obligados a comprometer estas

mismas parcelas en el siguiente ciclo agrícola, a la medida 1 “mejora y

mantenimiento del barbecho tradicional”, o bien, por segundo año repetir los

compromisos de la medida número 5

Compromisos generales para todas las medidas

Todas las parcelas acogidas al programa de medidas agroambientales,

independientemente de la medida a la que estén acogidos, deberán respetar unos

compromisos de carácter general para todas ellas.

1. No utilizar productos fitosanitarios sobre la parcela acogida.

2. No utilizar semillas tratadas o blindadas para la sementera.

3. No realizar quema de rastrojos.

4. No permitir el paso y pastoreo de ganado.

5. No se permite el uso ni el vertido de compost o de lodos de depuradora.

217

Todos los compromisos establecidos son de renovación anual, durante un periodo máximo

de 5 años

La dosis mínima de semilla que se recomienda para la sementera es de 180 Kg//ha

para el trigo, 170 Kg/ha para cebada, 190kg/ha para la veza y 120 kg/ha para otras

leguminosas.

PRIMAS COMPENSATORIAS

Los agricultores que decidan acogerse a este Programa recibirán una serie de primas

compensatorias, cuyas cantidades dependerán del tipo de medida que deseen aplicar (Tabla

1). Es aconsejable que los agricultores se acojan a las medidas de “Extensificación de la

producción agraria” ofrecidas por las administraciones autonómicas, que tienen un carácter

similar a las que ofrece el presente Programa, para así conseguir aumentar la cuantía de las

primas recibidas.

Tabla 1. Primas compensatorias según el tipo de medida

Prima

€/Ha/año

Medida 1 Mejora y mantenimiento del barbecho tradicional 138

Medida 2 Barbecho semillado con leguminosas 425

Medida 3 Retirada de la producción de tierras durante 4 años 287

Medida 4 Rotación de cultivos trigo-girasol 525

Medida 5 Cultivo de cereal no tratado 400

218

CONDICIONES DE CONCESION DE LAS AYUDAS

Para que las superficies puedan acogerse al programa de medidas agroambientales

deben de cumplir los siguientes requisitos:

Estar incluida dentro de alguno de los 7 polígonos descritos como zonas de

actuación del programa, estar situada en un área que por sus características sea susceptible

de ser ocupada por comunidades de aves esteparias y alejada más de 1000 metros de

núcleos de población, así como de carreteras asfaltadas, edificaciones habitadas, o cualquier

otra obra o lugar en que la actividad humana pueda causar molestias frecuentes a las aves.

DOCUMENTACIÓN A PARESENTAR POR LOS SOLICITANTES

Los solicitantes de las ayudas, una vez que hayan recibido en su domicilio una carta

con el documento de aceptación de las parcelas solicitadas, deben de presentar la siguiente

documentación:

1. Factura que acredite la compra de semillas que no contengan productos

fitosanitarios. En el caso de que la semilla sea propiedad del agricultor, deberá de

presentar una declaración jurada de no haber realizado ningún tratamiento a la

semilla destinada a la siembra de parcelas acogidas al programa.

2. En el caso de los agricultores que deseen acogerse a la medida 3, retirada de

producción de la tierra durante un periodo máximo de 5 años, deberán de entregar

una copia de una declaración o comprobante, que verifique el uso de la tierra a

labores agrícolas durante los últimos 3 años.

3. Fotocopia del DNI, N.I.F ó C.I.F.

4. Impreso de aceptación de parcelas incluidas en el programa, debidamente rellenado

y firmado.

219

PAGO DE LAS AYUDAS

Transcurrido el año agrícola y verificados los controles correspondientes, se enviará

por correo a cada agricultor una factura con la cantidad correspondiente a la ayuda. Una

vez firmada por el titular, éste deberá enviarla al Museo Nacional de Ciencias Naturales,

tras lo cual se procederá a tramitar el pago de las ayudas a los agricultores con la mayor

brevedad posible. Estos pagos serán efectuados en un pago único y por transferencia

bancaria.

CONTROLES Y SANCIONES POR INCUMPLIMIENTO DEL CONTRATRO

El control por parte de la empresa sobre el no uso de productos fitosanitarios,

abonos, calendario de labores y mantenimiento de barbechos y rastrojeras se realizará “in

situ” sobre el 100% de las parcelas acogidas al programa y será realizada numerosas veces a

lo largo del ciclo agrícola.

Cuando a través de los controles efectuados se comprueben anomalías en el

cumplimiento de las condiciones y compromisos suscritos en alguna de sus partes, la

empresa, en función de las circunstancias que concurran, podrá reducir las primas por

unidad de superficie durante el ciclo agrícola en transición, si las condiciones del programa

no han podido cumplirse en su totalidad, u optar por expulsar al propietario del programa

en el caso de que las condiciones y objetivos señalados en el programa dejasen de cumplirse

en su totalidad.

En el caso de incumplimiento de los compromisos establecidos en las medidas, se

aplicarán las siguientes sanciones:

1. Cuando se compruebe una variación en las normas de cultivo establecidas que

afecte entre el 5 y el 10 por ciento del total de la superficie de parcelas acogidas, se

procederá a reducir la prima total a percibir en dicho año por el agricultor, calculada

de acuerdo con las superficies reales, en un 20 por ciento.

220

2. Cuando la variación de la alternativa afecte entre el 10% y el 20% de la superficie, o

aun siendo inferior implique una disminución de la superficie de leguminosa, se

procederá a una disminución de la prima total a percibir en dicho año en un 50 por

ciento.

3. Cuando la variación supere el 20%, se procederá a anular la prima y a rescindir el

contrato

4. Cuando se compruebe reiteración durante más de un año en el incumplimiento de

alguno de los compromisos establecidos por el programa, se procederá a la

expulsión definitiva del propietario de las tierras del programa de medidas

agroambientales

FINANCIACIÓN

Henarsa S.A., concesionaria del Estado para la construcción y explotación de las

nuevas autopistas M-50 (tramo M-607 / N-IV, subtramo N-I / N-II) y R-2, a través del

Proyecto de medidas preventivas, correctoras y compensatorias de la afección de las

mencionadas infraestructuras a la población de avutardas y otras aves esteparias de la IBA

“Talamanca-Camarma”, y al LIC “Cuenca de los ríos Jarama y Henares”, elaborado por el

Museo Nacional de Ciencias Naturales, financiará en su totalidad las medidas que

componen este Programa.

221

Situación de las siete zonas donde se aplicará el Programa:

1. Ribatejada-Valdeolmos-Valdetorres de Jarama-Talamanca de Jarama- Valdepiélagos

2. Valdetorres de Jarama-Fuente el Saz

3. Camarma de Esteruelas-Daganzo de Arriba-Alcalá de Henares-Fresno de Torote-

Ribatejada

4. Cobeña-Paracuellos de Jarama

5. Ajalvir-Daganzo de Arriba

6. Camarma de Esteruelas-Meco

7. Cabanillas del Campo-Quer-Villanueva de la Torre

1

2

3

4

5

6

7

1

2

3

4

5

6

7